MiniMax-M2.5-mic

Slim model weights, executable usage and framework expectations in one repository.

• Contract manifest
• Transformers usage
• Transformers expectations
• Fixture provenance
• How the fixture was shrunk

Contract

MiniMax-M2.5 causal LM

Runnable usage and the framework expectations directory are linked above.

Slim model

Slim fixture: MiniMax-M2.5-mic

A 25.8M-parameter fixture for the MiniMaxM2 MIC, covering MiniMaxAI/MiniMax-M2.5. It is a test artifact, not a usable model, tagged test-artifact to keep it out of candidate-model search.

• Fixture key caec3890e5c1. It covers every checkpoint whose processor files are byte-identical and whose config matches after shrinking.
• Same architecture, smaller config: MiniMaxM2ForCausalLM, shrunk by the generic rule in sweep.py (num_hidden_layers 2).
• Processor files copied byte for byte from MiniMaxAI/MiniMax-M2.5: LICENSE-MODEL, chat_template.jinja, configuration_minimax_m2.py, generation_config.json, merges.txt, modeling_minimax_m2.py, tokenizer.json, tokenizer_config.json, vocab.json.
• Random weights (seed 17). They passed the input-sensitivity gate: every perturbation of the contract's inputs moved the logits by at least 0.001 (weakest: input_ids:roll_batch, 0.635), so no training was needed.
• Same tensor names as the real checkpoint. The real checkpoint has every parameter, and so does the fixture.
• Saved in float32, the dtype MiniMaxAI/MiniMax-M2.5 loads in by default.
• Finite in float32, bfloat16, and float16 (forward, loss, gradients).

Built by build_fixtures.py; build.json records the pinned sources, gate results, and file hashes. This repository includes the slim weights, usage and framework expectations; rebuild the other files with python example/build_fixtures.py --fixture MiniMax-M2.5-mic.