Qwen3-8B — official Apple Core AI export
Pre-converted .aimodel bundles from Apple's official
coreai-models export recipe — unmodified,
with the exact environment, hashes, and measured performance published.
uv run coreai.llm.export qwen3-8b
Why pre-converted bundles?
- The conversion needs a big-RAM Mac (the 20B export was done on 128 GB); running only needs enough RAM to mmap the artifact.
- An
.aimodelis a build artifact, not a pure function of the recipe — the same export command produced a 2.2× slower artifact across the macOS 26 → 27β boundary (forensics). Hosted artifacts + hashes are the reproducible ground truth; every bundle here is exactly the one measured in apple-silicon-llm-bench.
Bundles & integrity
| Bundle | Contents | SHA-256 (main.mlirb) |
|---|---|---|
macos/ |
macOS dynamic, int4 | f659250441d88f9eaf6f260b11e2644edac9245b7bea89e30c70dc1960ef953b |
Measured (Apple's official llm-benchmark, greedy)
| Bundle | Protocol | Decode tok/s | Prefill | Load (warm) | Peak RSS |
|---|---|---|---|---|---|
| macos | M4 Max, 512p/1024g | 94.1 | 912 | 0.64 s | 9.3 GB |
Export environment
- macOS 27.0 beta (build 26A5353q) · Xcode 27.0 (27A5194q)
coreai-core 1.0.0b1·coreai-torch 0.4.0·coreai-opt 0.2.0·torch 2.9.0- apple/coreai-models @
b1cb71b(export code identical to upstream0c1055f)
Run it
# CLI (from a coreai-models checkout)
swift run -c release llm-runner --model <downloaded-bundle-dir> --prompt "Hello"
swift run -c release llm-benchmark --model <downloaded-bundle-dir>
Or chat with it in CoreAIChatMac (point "Choose Models Folder…" at the download directory).
iOS static bundles must be AOT-compiled before device use:
xcrun coreai-build compile <ir>.aimodel --platform iOS --preferred-compute neural-engine --architecture h18p
(h18p = iPhone 17 Pro), then set metadata.json assets.main to the .aimodelc.
Maintained alongside coreai-model-zoo (community models) and coreai-samples (apps).
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support