Datasets:
record_id stringlengths 105 151 | provider_id stringclasses 2
values | model_id stringclasses 9
values | offering_id stringclasses 9
values | api_model_id stringclasses 9
values | model_name stringclasses 9
values | region stringclasses 4
values | service_tier stringclasses 22
values | billing_mode stringclasses 4
values | price_component stringclasses 9
values | context_min_tokens float64 0 272k ⌀ | context_max_tokens float64 272k 1.05M ⌀ | native_amount float64 0 60 | currency stringclasses 1
value | native_unit stringclasses 5
values | normalized_usd_per_1m_tokens float64 0.01 60 ⌀ | observed_at stringclasses 1
value | source_ids stringclasses 9
values | notes stringclasses 20
values |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
kapi:price-record:anthropic:claude-fable-5-global-routing-supported-countries-batch-50-percent-batch-cache-write-1h-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | global_routing_supported_countries | batch_50_percent | batch | cache_write_1h | 0 | 1,000,000 | 10 | USD | per_1m_tokens | 10 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-global-routing-supported-countries-batch-50-percent-batch-cache-write-5m-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | global_routing_supported_countries | batch_50_percent | batch | cache_write_5m | 0 | 1,000,000 | 6.25 | USD | per_1m_tokens | 6.25 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-global-routing-supported-countries-batch-50-percent-batch-cached-input-read-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | global_routing_supported_countries | batch_50_percent | batch | cached_input_read | 0 | 1,000,000 | 0.5 | USD | per_1m_tokens | 0.5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-global-routing-supported-countries-batch-50-percent-batch-input-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | global_routing_supported_countries | batch_50_percent | batch | input | 0 | 1,000,000 | 5 | USD | per_1m_tokens | 5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-global-routing-supported-countries-batch-50-percent-batch-output-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | global_routing_supported_countries | batch_50_percent | batch | output | 0 | 1,000,000 | 25 | USD | per_1m_tokens | 25 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-global-routing-supported-countries-standard-on-demand-on-demand-cache-write-1h-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | global_routing_supported_countries | standard_on_demand | on_demand | cache_write_1h | 0 | 1,000,000 | 20 | USD | per_1m_tokens | 20 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-global-routing-supported-countries-standard-on-demand-on-demand-cache-write-5m-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | global_routing_supported_countries | standard_on_demand | on_demand | cache_write_5m | 0 | 1,000,000 | 12.5 | USD | per_1m_tokens | 12.5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-global-routing-supported-countries-standard-on-demand-on-demand-cached-input-read-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | global_routing_supported_countries | standard_on_demand | on_demand | cached_input_read | 0 | 1,000,000 | 1 | USD | per_1m_tokens | 1 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-global-routing-supported-countries-standard-on-demand-on-demand-input-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | global_routing_supported_countries | standard_on_demand | on_demand | input | 0 | 1,000,000 | 10 | USD | per_1m_tokens | 10 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-global-routing-supported-countries-standard-on-demand-on-demand-output-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | global_routing_supported_countries | standard_on_demand | on_demand | output | 0 | 1,000,000 | 50 | USD | per_1m_tokens | 50 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-us-only-inference-batch-50-percent-1-1x-us-batch-cache-write-1h-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | us_only_inference | batch_50_percent_1_1x_us | batch | cache_write_1h | 0 | 1,000,000 | 11 | USD | per_1m_tokens | 11 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-us-only-inference-batch-50-percent-1-1x-us-batch-cache-write-5m-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | us_only_inference | batch_50_percent_1_1x_us | batch | cache_write_5m | 0 | 1,000,000 | 6.875 | USD | per_1m_tokens | 6.875 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-us-only-inference-batch-50-percent-1-1x-us-batch-cached-input-read-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | us_only_inference | batch_50_percent_1_1x_us | batch | cached_input_read | 0 | 1,000,000 | 0.55 | USD | per_1m_tokens | 0.55 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-us-only-inference-batch-50-percent-1-1x-us-batch-input-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | us_only_inference | batch_50_percent_1_1x_us | batch | input | 0 | 1,000,000 | 5.5 | USD | per_1m_tokens | 5.5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-us-only-inference-batch-50-percent-1-1x-us-batch-output-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | us_only_inference | batch_50_percent_1_1x_us | batch | output | 0 | 1,000,000 | 27.5 | USD | per_1m_tokens | 27.5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-cache-write-1h-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | cache_write_1h | 0 | 1,000,000 | 22 | USD | per_1m_tokens | 22 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-cache-write-5m-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | cache_write_5m | 0 | 1,000,000 | 13.75 | USD | per_1m_tokens | 13.75 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-cached-input-read-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | cached_input_read | 0 | 1,000,000 | 1.1 | USD | per_1m_tokens | 1.1 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-input-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | input | 0 | 1,000,000 | 11 | USD | per_1m_tokens | 11 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-fable-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-output-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-fable-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-fable-5 | claude-fable-5 | Claude Fable 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | output | 0 | 1,000,000 | 55 | USD | per_1m_tokens | 55 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-global-routing-supported-countries-batch-50-percent-batch-cache-write-1h-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | global_routing_supported_countries | batch_50_percent | batch | cache_write_1h | 0 | 1,000,000 | 10 | USD | per_1m_tokens | 10 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-global-routing-supported-countries-batch-50-percent-batch-cache-write-5m-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | global_routing_supported_countries | batch_50_percent | batch | cache_write_5m | 0 | 1,000,000 | 6.25 | USD | per_1m_tokens | 6.25 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-global-routing-supported-countries-batch-50-percent-batch-cached-input-read-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | global_routing_supported_countries | batch_50_percent | batch | cached_input_read | 0 | 1,000,000 | 0.5 | USD | per_1m_tokens | 0.5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-global-routing-supported-countries-batch-50-percent-batch-input-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | global_routing_supported_countries | batch_50_percent | batch | input | 0 | 1,000,000 | 5 | USD | per_1m_tokens | 5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-global-routing-supported-countries-batch-50-percent-batch-output-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | global_routing_supported_countries | batch_50_percent | batch | output | 0 | 1,000,000 | 25 | USD | per_1m_tokens | 25 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-global-routing-supported-countries-standard-on-demand-on-demand-cache-write-1h-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | global_routing_supported_countries | standard_on_demand | on_demand | cache_write_1h | 0 | 1,000,000 | 20 | USD | per_1m_tokens | 20 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-global-routing-supported-countries-standard-on-demand-on-demand-cache-write-5m-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | global_routing_supported_countries | standard_on_demand | on_demand | cache_write_5m | 0 | 1,000,000 | 12.5 | USD | per_1m_tokens | 12.5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-global-routing-supported-countries-standard-on-demand-on-demand-cached-input-read-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | global_routing_supported_countries | standard_on_demand | on_demand | cached_input_read | 0 | 1,000,000 | 1 | USD | per_1m_tokens | 1 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-global-routing-supported-countries-standard-on-demand-on-demand-input-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | global_routing_supported_countries | standard_on_demand | on_demand | input | 0 | 1,000,000 | 10 | USD | per_1m_tokens | 10 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-global-routing-supported-countries-standard-on-demand-on-demand-output-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | global_routing_supported_countries | standard_on_demand | on_demand | output | 0 | 1,000,000 | 50 | USD | per_1m_tokens | 50 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-us-only-inference-batch-50-percent-1-1x-us-batch-cache-write-1h-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | us_only_inference | batch_50_percent_1_1x_us | batch | cache_write_1h | 0 | 1,000,000 | 11 | USD | per_1m_tokens | 11 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-us-only-inference-batch-50-percent-1-1x-us-batch-cache-write-5m-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | us_only_inference | batch_50_percent_1_1x_us | batch | cache_write_5m | 0 | 1,000,000 | 6.875 | USD | per_1m_tokens | 6.875 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-us-only-inference-batch-50-percent-1-1x-us-batch-cached-input-read-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | us_only_inference | batch_50_percent_1_1x_us | batch | cached_input_read | 0 | 1,000,000 | 0.55 | USD | per_1m_tokens | 0.55 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-us-only-inference-batch-50-percent-1-1x-us-batch-input-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | us_only_inference | batch_50_percent_1_1x_us | batch | input | 0 | 1,000,000 | 5.5 | USD | per_1m_tokens | 5.5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-us-only-inference-batch-50-percent-1-1x-us-batch-output-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | us_only_inference | batch_50_percent_1_1x_us | batch | output | 0 | 1,000,000 | 27.5 | USD | per_1m_tokens | 27.5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-cache-write-1h-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | cache_write_1h | 0 | 1,000,000 | 22 | USD | per_1m_tokens | 22 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-cache-write-5m-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | cache_write_5m | 0 | 1,000,000 | 13.75 | USD | per_1m_tokens | 13.75 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-cached-input-read-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | cached_input_read | 0 | 1,000,000 | 1.1 | USD | per_1m_tokens | 1.1 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-input-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | input | 0 | 1,000,000 | 11 | USD | per_1m_tokens | 11 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-mythos-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-output-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-mythos-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-mythos-5 | claude-mythos-5 | Claude Mythos 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | output | 0 | 1,000,000 | 55 | USD | per_1m_tokens | 55 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-global-routing-supported-countries-batch-50-percent-batch-cache-write-1h-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | global_routing_supported_countries | batch_50_percent | batch | cache_write_1h | 0 | 1,000,000 | 5 | USD | per_1m_tokens | 5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-global-routing-supported-countries-batch-50-percent-batch-cache-write-5m-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | global_routing_supported_countries | batch_50_percent | batch | cache_write_5m | 0 | 1,000,000 | 3.125 | USD | per_1m_tokens | 3.125 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-global-routing-supported-countries-batch-50-percent-batch-cached-input-read-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | global_routing_supported_countries | batch_50_percent | batch | cached_input_read | 0 | 1,000,000 | 0.25 | USD | per_1m_tokens | 0.25 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-global-routing-supported-countries-batch-50-percent-batch-input-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | global_routing_supported_countries | batch_50_percent | batch | input | 0 | 1,000,000 | 2.5 | USD | per_1m_tokens | 2.5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-global-routing-supported-countries-batch-50-percent-batch-output-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | global_routing_supported_countries | batch_50_percent | batch | output | 0 | 1,000,000 | 12.5 | USD | per_1m_tokens | 12.5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-global-routing-supported-countries-fast-direct-api-priority-input-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | global_routing_supported_countries | fast_direct_api | priority | input | 0 | 1,000,000 | 10 | USD | per_1m_tokens | 10 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23 | Fast mode is limited to the direct Claude API and is not available with Message Batches. Current published Fast pricing is $10 input / $50 output per 1M tokens. |
kapi:price-record:anthropic:claude-opus-5-global-routing-supported-countries-fast-direct-api-priority-output-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | global_routing_supported_countries | fast_direct_api | priority | output | 0 | 1,000,000 | 50 | USD | per_1m_tokens | 50 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23 | Fast mode is limited to the direct Claude API and is not available with Message Batches. Current published Fast pricing is $10 input / $50 output per 1M tokens. |
kapi:price-record:anthropic:claude-opus-5-global-routing-supported-countries-standard-on-demand-on-demand-cache-write-1h-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | global_routing_supported_countries | standard_on_demand | on_demand | cache_write_1h | 0 | 1,000,000 | 10 | USD | per_1m_tokens | 10 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-global-routing-supported-countries-standard-on-demand-on-demand-cache-write-5m-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | global_routing_supported_countries | standard_on_demand | on_demand | cache_write_5m | 0 | 1,000,000 | 6.25 | USD | per_1m_tokens | 6.25 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-global-routing-supported-countries-standard-on-demand-on-demand-cached-input-read-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | global_routing_supported_countries | standard_on_demand | on_demand | cached_input_read | 0 | 1,000,000 | 0.5 | USD | per_1m_tokens | 0.5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-global-routing-supported-countries-standard-on-demand-on-demand-input-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | global_routing_supported_countries | standard_on_demand | on_demand | input | 0 | 1,000,000 | 5 | USD | per_1m_tokens | 5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-global-routing-supported-countries-standard-on-demand-on-demand-output-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | global_routing_supported_countries | standard_on_demand | on_demand | output | 0 | 1,000,000 | 25 | USD | per_1m_tokens | 25 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-us-only-inference-batch-50-percent-1-1x-us-batch-cache-write-1h-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | us_only_inference | batch_50_percent_1_1x_us | batch | cache_write_1h | 0 | 1,000,000 | 5.5 | USD | per_1m_tokens | 5.5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-us-only-inference-batch-50-percent-1-1x-us-batch-cache-write-5m-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | us_only_inference | batch_50_percent_1_1x_us | batch | cache_write_5m | 0 | 1,000,000 | 3.4375 | USD | per_1m_tokens | 3.4375 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-us-only-inference-batch-50-percent-1-1x-us-batch-cached-input-read-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | us_only_inference | batch_50_percent_1_1x_us | batch | cached_input_read | 0 | 1,000,000 | 0.275 | USD | per_1m_tokens | 0.275 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-us-only-inference-batch-50-percent-1-1x-us-batch-input-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | us_only_inference | batch_50_percent_1_1x_us | batch | input | 0 | 1,000,000 | 2.75 | USD | per_1m_tokens | 2.75 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-us-only-inference-batch-50-percent-1-1x-us-batch-output-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | us_only_inference | batch_50_percent_1_1x_us | batch | output | 0 | 1,000,000 | 13.75 | USD | per_1m_tokens | 13.75 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-cache-write-1h-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | cache_write_1h | 0 | 1,000,000 | 11 | USD | per_1m_tokens | 11 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-cache-write-5m-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | cache_write_5m | 0 | 1,000,000 | 6.875 | USD | per_1m_tokens | 6.875 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-cached-input-read-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | cached_input_read | 0 | 1,000,000 | 0.55 | USD | per_1m_tokens | 0.55 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-input-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | input | 0 | 1,000,000 | 5.5 | USD | per_1m_tokens | 5.5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-opus-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-output-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-opus-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-opus-5 | claude-opus-5 | Claude Opus 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | output | 0 | 1,000,000 | 27.5 | USD | per_1m_tokens | 27.5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-platform-tools-global-routing-supported-countries-code-execution-on-demand-tool-call-na-na | kng:provider:anthropic | kng:anthropic.claude-platform-tools | kapi:offering:anthropic:first-party-api:anthropic-claude-platform-tools | claude-platform-tools | Claude platform tools | global_routing_supported_countries | code_execution | on_demand | tool_call | null | null | 0.05 | USD | per_hour | null | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23 | Standalone code execution has a 5-minute minimum; the first 1,550 organization hours each month are free, then $0.05 per container-hour. No separate compute charge when paired with web search or web fetch, excluding token charges. |
kapi:price-record:anthropic:claude-platform-tools-global-routing-supported-countries-web-fetch-on-demand-tool-call-na-na | kng:provider:anthropic | kng:anthropic.claude-platform-tools | kapi:offering:anthropic:first-party-api:anthropic-claude-platform-tools | claude-platform-tools | Claude platform tools | global_routing_supported_countries | web_fetch | on_demand | tool_call | null | null | 0 | USD | per_call | null | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23 | Web fetch has no per-call fee, but fetched content is billed as input tokens. |
kapi:price-record:anthropic:claude-platform-tools-global-routing-supported-countries-web-search-on-demand-search-call-na-na | kng:provider:anthropic | kng:anthropic.claude-platform-tools | kapi:offering:anthropic:first-party-api:anthropic-claude-platform-tools | claude-platform-tools | Claude platform tools | global_routing_supported_countries | web_search | on_demand | search_call | null | null | 0.01 | USD | per_call | null | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23 | Web search costs $10 per 1,000 searches; tokens used by search are billed separately. |
kapi:price-record:anthropic:claude-sonnet-5-global-routing-supported-countries-batch-50-percent-batch-cache-write-1h-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | global_routing_supported_countries | batch_50_percent | batch | cache_write_1h | 0 | 1,000,000 | 2 | USD | per_1m_tokens | 2 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-global-routing-supported-countries-batch-50-percent-batch-cache-write-5m-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | global_routing_supported_countries | batch_50_percent | batch | cache_write_5m | 0 | 1,000,000 | 1.25 | USD | per_1m_tokens | 1.25 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-global-routing-supported-countries-batch-50-percent-batch-cached-input-read-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | global_routing_supported_countries | batch_50_percent | batch | cached_input_read | 0 | 1,000,000 | 0.1 | USD | per_1m_tokens | 0.1 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-global-routing-supported-countries-batch-50-percent-batch-input-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | global_routing_supported_countries | batch_50_percent | batch | input | 0 | 1,000,000 | 1 | USD | per_1m_tokens | 1 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-global-routing-supported-countries-batch-50-percent-batch-output-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | global_routing_supported_countries | batch_50_percent | batch | output | 0 | 1,000,000 | 5 | USD | per_1m_tokens | 5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-global-routing-supported-countries-standard-on-demand-on-demand-cache-write-1h-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | global_routing_supported_countries | standard_on_demand | on_demand | cache_write_1h | 0 | 1,000,000 | 4 | USD | per_1m_tokens | 4 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-global-routing-supported-countries-standard-on-demand-on-demand-cache-write-5m-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | global_routing_supported_countries | standard_on_demand | on_demand | cache_write_5m | 0 | 1,000,000 | 2.5 | USD | per_1m_tokens | 2.5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-global-routing-supported-countries-standard-on-demand-on-demand-cached-input-read-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | global_routing_supported_countries | standard_on_demand | on_demand | cached_input_read | 0 | 1,000,000 | 0.2 | USD | per_1m_tokens | 0.2 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-global-routing-supported-countries-standard-on-demand-on-demand-input-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | global_routing_supported_countries | standard_on_demand | on_demand | input | 0 | 1,000,000 | 2 | USD | per_1m_tokens | 2 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-global-routing-supported-countries-standard-on-demand-on-demand-output-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | global_routing_supported_countries | standard_on_demand | on_demand | output | 0 | 1,000,000 | 10 | USD | per_1m_tokens | 10 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Global inference is the direct API default. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-us-only-inference-batch-50-percent-1-1x-us-batch-cache-write-1h-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | us_only_inference | batch_50_percent_1_1x_us | batch | cache_write_1h | 0 | 1,000,000 | 2.2 | USD | per_1m_tokens | 2.2 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-us-only-inference-batch-50-percent-1-1x-us-batch-cache-write-5m-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | us_only_inference | batch_50_percent_1_1x_us | batch | cache_write_5m | 0 | 1,000,000 | 1.375 | USD | per_1m_tokens | 1.375 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-us-only-inference-batch-50-percent-1-1x-us-batch-cached-input-read-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | us_only_inference | batch_50_percent_1_1x_us | batch | cached_input_read | 0 | 1,000,000 | 0.11 | USD | per_1m_tokens | 0.11 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-us-only-inference-batch-50-percent-1-1x-us-batch-input-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | us_only_inference | batch_50_percent_1_1x_us | batch | input | 0 | 1,000,000 | 1.1 | USD | per_1m_tokens | 1.1 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-us-only-inference-batch-50-percent-1-1x-us-batch-output-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | us_only_inference | batch_50_percent_1_1x_us | batch | output | 0 | 1,000,000 | 5.5 | USD | per_1m_tokens | 5.5 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23|kapi:source:anthropic:batch-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-cache-write-1h-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | cache_write_1h | 0 | 1,000,000 | 4.4 | USD | per_1m_tokens | 4.4 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-cache-write-5m-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | cache_write_5m | 0 | 1,000,000 | 2.75 | USD | per_1m_tokens | 2.75 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-cached-input-read-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | cached_input_read | 0 | 1,000,000 | 0.22 | USD | per_1m_tokens | 0.22 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-input-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | input | 0 | 1,000,000 | 2.2 | USD | per_1m_tokens | 2.2 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:anthropic:claude-sonnet-5-us-only-inference-standard-on-demand-1-1x-us-on-demand-output-0-1000000 | kng:provider:anthropic | kng:anthropic.claude-sonnet-5 | kapi:offering:anthropic:first-party-api:anthropic-claude-sonnet-5 | claude-sonnet-5 | Claude Sonnet 5 | us_only_inference | standard_on_demand_1_1x_us | on_demand | output | 0 | 1,000,000 | 11 | USD | per_1m_tokens | 11 | 2026-08-24T04:11:22.000Z | kapi:source:anthropic:pricing-2026-08-23|kapi:source:anthropic:models-2026-08-23|kapi:source:anthropic:caching-2026-08-23 | Claude 5 supports up to 1M input context at the same listed token rate. Direct API US-only inference uses the documented 1.1× multiplier. Claude 4.7+ uses a newer tokenizer that may produce about 30% more tokens for the same text; actual change is workload dependent. |
kapi:price-record:openai:gpt-5-6-luna-global-default-processing-batch-50-percent-long-context-batch-cache-write-272001-1050000 | kng:provider:openai | kng:openai.gpt-5.6-luna | kapi:offering:openai:first-party-api:openai-gpt-5-6-luna | gpt-5.6-luna | GPT-5.6 Luna | global_default_processing | batch_50_percent_long_context | batch | cache_write | 272,001 | 1,050,000 | 0.25 | USD | per_1m_tokens | 0.25 | 2026-08-24T04:11:22.000Z | kapi:source:openai:pricing-2026-08-23|kapi:source:openai:models-2026-08-23|kapi:source:openai:caching-2026-08-23|kapi:source:openai:batch-2026-08-23 | POST /v1/responses with gpt-5.6-luna. When input exceeds 272K tokens, the long-context rate applies to the entire request: 2× input-side and 1.5× output. |
kapi:price-record:openai:gpt-5-6-luna-global-default-processing-batch-50-percent-long-context-batch-cached-input-read-272001-1050000 | kng:provider:openai | kng:openai.gpt-5.6-luna | kapi:offering:openai:first-party-api:openai-gpt-5-6-luna | gpt-5.6-luna | GPT-5.6 Luna | global_default_processing | batch_50_percent_long_context | batch | cached_input_read | 272,001 | 1,050,000 | 0.02 | USD | per_1m_tokens | 0.02 | 2026-08-24T04:11:22.000Z | kapi:source:openai:pricing-2026-08-23|kapi:source:openai:models-2026-08-23|kapi:source:openai:caching-2026-08-23|kapi:source:openai:batch-2026-08-23 | POST /v1/responses with gpt-5.6-luna. When input exceeds 272K tokens, the long-context rate applies to the entire request: 2× input-side and 1.5× output. |
kapi:price-record:openai:gpt-5-6-luna-global-default-processing-batch-50-percent-long-context-batch-input-272001-1050000 | kng:provider:openai | kng:openai.gpt-5.6-luna | kapi:offering:openai:first-party-api:openai-gpt-5-6-luna | gpt-5.6-luna | GPT-5.6 Luna | global_default_processing | batch_50_percent_long_context | batch | input | 272,001 | 1,050,000 | 0.2 | USD | per_1m_tokens | 0.2 | 2026-08-24T04:11:22.000Z | kapi:source:openai:pricing-2026-08-23|kapi:source:openai:models-2026-08-23|kapi:source:openai:caching-2026-08-23|kapi:source:openai:batch-2026-08-23 | POST /v1/responses with gpt-5.6-luna. When input exceeds 272K tokens, the long-context rate applies to the entire request: 2× input-side and 1.5× output. |
kapi:price-record:openai:gpt-5-6-luna-global-default-processing-batch-50-percent-long-context-batch-output-272001-1050000 | kng:provider:openai | kng:openai.gpt-5.6-luna | kapi:offering:openai:first-party-api:openai-gpt-5-6-luna | gpt-5.6-luna | GPT-5.6 Luna | global_default_processing | batch_50_percent_long_context | batch | output | 272,001 | 1,050,000 | 0.9 | USD | per_1m_tokens | 0.9 | 2026-08-24T04:11:22.000Z | kapi:source:openai:pricing-2026-08-23|kapi:source:openai:models-2026-08-23|kapi:source:openai:caching-2026-08-23|kapi:source:openai:batch-2026-08-23 | POST /v1/responses with gpt-5.6-luna. When input exceeds 272K tokens, the long-context rate applies to the entire request: 2× input-side and 1.5× output. |
kapi:price-record:openai:gpt-5-6-luna-global-default-processing-batch-50-percent-short-context-batch-cache-write-0-272000 | kng:provider:openai | kng:openai.gpt-5.6-luna | kapi:offering:openai:first-party-api:openai-gpt-5-6-luna | gpt-5.6-luna | GPT-5.6 Luna | global_default_processing | batch_50_percent_short_context | batch | cache_write | 0 | 272,000 | 0.125 | USD | per_1m_tokens | 0.125 | 2026-08-24T04:11:22.000Z | kapi:source:openai:pricing-2026-08-23|kapi:source:openai:models-2026-08-23|kapi:source:openai:caching-2026-08-23|kapi:source:openai:batch-2026-08-23 | POST /v1/responses with gpt-5.6-luna. Short-context band covers requests with at most 272K input tokens. |
kapi:price-record:openai:gpt-5-6-luna-global-default-processing-batch-50-percent-short-context-batch-cached-input-read-0-272000 | kng:provider:openai | kng:openai.gpt-5.6-luna | kapi:offering:openai:first-party-api:openai-gpt-5-6-luna | gpt-5.6-luna | GPT-5.6 Luna | global_default_processing | batch_50_percent_short_context | batch | cached_input_read | 0 | 272,000 | 0.01 | USD | per_1m_tokens | 0.01 | 2026-08-24T04:11:22.000Z | kapi:source:openai:pricing-2026-08-23|kapi:source:openai:models-2026-08-23|kapi:source:openai:caching-2026-08-23|kapi:source:openai:batch-2026-08-23 | POST /v1/responses with gpt-5.6-luna. Short-context band covers requests with at most 272K input tokens. |
kapi:price-record:openai:gpt-5-6-luna-global-default-processing-batch-50-percent-short-context-batch-input-0-272000 | kng:provider:openai | kng:openai.gpt-5.6-luna | kapi:offering:openai:first-party-api:openai-gpt-5-6-luna | gpt-5.6-luna | GPT-5.6 Luna | global_default_processing | batch_50_percent_short_context | batch | input | 0 | 272,000 | 0.1 | USD | per_1m_tokens | 0.1 | 2026-08-24T04:11:22.000Z | kapi:source:openai:pricing-2026-08-23|kapi:source:openai:models-2026-08-23|kapi:source:openai:caching-2026-08-23|kapi:source:openai:batch-2026-08-23 | POST /v1/responses with gpt-5.6-luna. Short-context band covers requests with at most 272K input tokens. |
kapi:price-record:openai:gpt-5-6-luna-global-default-processing-batch-50-percent-short-context-batch-output-0-272000 | kng:provider:openai | kng:openai.gpt-5.6-luna | kapi:offering:openai:first-party-api:openai-gpt-5-6-luna | gpt-5.6-luna | GPT-5.6 Luna | global_default_processing | batch_50_percent_short_context | batch | output | 0 | 272,000 | 0.6 | USD | per_1m_tokens | 0.6 | 2026-08-24T04:11:22.000Z | kapi:source:openai:pricing-2026-08-23|kapi:source:openai:models-2026-08-23|kapi:source:openai:caching-2026-08-23|kapi:source:openai:batch-2026-08-23 | POST /v1/responses with gpt-5.6-luna. Short-context band covers requests with at most 272K input tokens. |
kapi:price-record:openai:gpt-5-6-luna-global-default-processing-fast-2x-standard-long-context-priority-cache-write-272001-1050000 | kng:provider:openai | kng:openai.gpt-5.6-luna | kapi:offering:openai:first-party-api:openai-gpt-5-6-luna | gpt-5.6-luna | GPT-5.6 Luna | global_default_processing | fast_2x_standard_long_context | priority | cache_write | 272,001 | 1,050,000 | 1 | USD | per_1m_tokens | 1 | 2026-08-24T04:11:22.000Z | kapi:source:openai:pricing-2026-08-23|kapi:source:openai:models-2026-08-23|kapi:source:openai:caching-2026-08-23|kapi:source:openai:fast-2026-08-23 | POST /v1/responses with gpt-5.6-luna. When input exceeds 272K tokens, the long-context rate applies to the entire request: 2× input-side and 1.5× output. |
kapi:price-record:openai:gpt-5-6-luna-global-default-processing-fast-2x-standard-long-context-priority-cached-input-read-272001-1050000 | kng:provider:openai | kng:openai.gpt-5.6-luna | kapi:offering:openai:first-party-api:openai-gpt-5-6-luna | gpt-5.6-luna | GPT-5.6 Luna | global_default_processing | fast_2x_standard_long_context | priority | cached_input_read | 272,001 | 1,050,000 | 0.08 | USD | per_1m_tokens | 0.08 | 2026-08-24T04:11:22.000Z | kapi:source:openai:pricing-2026-08-23|kapi:source:openai:models-2026-08-23|kapi:source:openai:caching-2026-08-23|kapi:source:openai:fast-2026-08-23 | POST /v1/responses with gpt-5.6-luna. When input exceeds 272K tokens, the long-context rate applies to the entire request: 2× input-side and 1.5× output. |
kapi:price-record:openai:gpt-5-6-luna-global-default-processing-fast-2x-standard-long-context-priority-input-272001-1050000 | kng:provider:openai | kng:openai.gpt-5.6-luna | kapi:offering:openai:first-party-api:openai-gpt-5-6-luna | gpt-5.6-luna | GPT-5.6 Luna | global_default_processing | fast_2x_standard_long_context | priority | input | 272,001 | 1,050,000 | 0.8 | USD | per_1m_tokens | 0.8 | 2026-08-24T04:11:22.000Z | kapi:source:openai:pricing-2026-08-23|kapi:source:openai:models-2026-08-23|kapi:source:openai:caching-2026-08-23|kapi:source:openai:fast-2026-08-23 | POST /v1/responses with gpt-5.6-luna. When input exceeds 272K tokens, the long-context rate applies to the entire request: 2× input-side and 1.5× output. |
kapi:price-record:openai:gpt-5-6-luna-global-default-processing-fast-2x-standard-long-context-priority-output-272001-1050000 | kng:provider:openai | kng:openai.gpt-5.6-luna | kapi:offering:openai:first-party-api:openai-gpt-5-6-luna | gpt-5.6-luna | GPT-5.6 Luna | global_default_processing | fast_2x_standard_long_context | priority | output | 272,001 | 1,050,000 | 3.6 | USD | per_1m_tokens | 3.6 | 2026-08-24T04:11:22.000Z | kapi:source:openai:pricing-2026-08-23|kapi:source:openai:models-2026-08-23|kapi:source:openai:caching-2026-08-23|kapi:source:openai:fast-2026-08-23 | POST /v1/responses with gpt-5.6-luna. When input exceeds 272K tokens, the long-context rate applies to the entire request: 2× input-side and 1.5× output. |
kapi:price-record:openai:gpt-5-6-luna-global-default-processing-fast-2x-standard-short-context-priority-cache-write-0-272000 | kng:provider:openai | kng:openai.gpt-5.6-luna | kapi:offering:openai:first-party-api:openai-gpt-5-6-luna | gpt-5.6-luna | GPT-5.6 Luna | global_default_processing | fast_2x_standard_short_context | priority | cache_write | 0 | 272,000 | 0.5 | USD | per_1m_tokens | 0.5 | 2026-08-24T04:11:22.000Z | kapi:source:openai:pricing-2026-08-23|kapi:source:openai:models-2026-08-23|kapi:source:openai:caching-2026-08-23|kapi:source:openai:fast-2026-08-23 | POST /v1/responses with gpt-5.6-luna. Short-context band covers requests with at most 272K input tokens. |
kapi:price-record:openai:gpt-5-6-luna-global-default-processing-fast-2x-standard-short-context-priority-cached-input-read-0-272000 | kng:provider:openai | kng:openai.gpt-5.6-luna | kapi:offering:openai:first-party-api:openai-gpt-5-6-luna | gpt-5.6-luna | GPT-5.6 Luna | global_default_processing | fast_2x_standard_short_context | priority | cached_input_read | 0 | 272,000 | 0.04 | USD | per_1m_tokens | 0.04 | 2026-08-24T04:11:22.000Z | kapi:source:openai:pricing-2026-08-23|kapi:source:openai:models-2026-08-23|kapi:source:openai:caching-2026-08-23|kapi:source:openai:fast-2026-08-23 | POST /v1/responses with gpt-5.6-luna. Short-context band covers requests with at most 272K input tokens. |
kapi:price-record:openai:gpt-5-6-luna-global-default-processing-fast-2x-standard-short-context-priority-input-0-272000 | kng:provider:openai | kng:openai.gpt-5.6-luna | kapi:offering:openai:first-party-api:openai-gpt-5-6-luna | gpt-5.6-luna | GPT-5.6 Luna | global_default_processing | fast_2x_standard_short_context | priority | input | 0 | 272,000 | 0.4 | USD | per_1m_tokens | 0.4 | 2026-08-24T04:11:22.000Z | kapi:source:openai:pricing-2026-08-23|kapi:source:openai:models-2026-08-23|kapi:source:openai:caching-2026-08-23|kapi:source:openai:fast-2026-08-23 | POST /v1/responses with gpt-5.6-luna. Short-context band covers requests with at most 272K input tokens. |
End of preview. Expand in Data Studio
OpenAI and Anthropic API Price Index — August 2026
A versioned research release containing normalized OpenAI and Anthropic API price records, source identifiers, methodology, machine-readable metadata, and a reusable provider-level chart bundle.
Dataset structure
The 212 normalized price records in the focused nine-offering OpenAI and Anthropic refresh.
Variables: provider_id, model_id, price_component, native_amount, native_unit, normalized_usd_per_1m_tokens, observed_at, source_ids.
Provenance and methodology
See methodology.md and provenance/sources.csv in the release package.
Intended uses
Use for transparent analysis, comparison, and reproducible charting within the scope documented in the methodology.
Limitations and bias
- Not a market census
- Prices and product terms can change
Creators
- Curtis Pyke (Conceptualization; Data curation; Formal analysis; Methodology; Validation; Visualization; Writing – original draft)
Citation
Use CITATION.cff. Cite the exact-version DOI when available; the concept DOI refers to the evolving product.
Independence
Independently funded. See DISCLOSURE.md for the complete statement.
- Downloads last month
- 26