MTPLX
Collection
3 items • Updated
Qwen3.5-9B-MTPLX
MTPLX-branded multi-token-prediction model for Apple Silicon (MLX).
Forged with MTPLX Forge from
qwen/qwen3.5-9b.
- 8-bit quantization
- Group 32
- Calibrated native-MTP draft head
Verification
- Best depth: D3
- Multiplier vs autoregressive baseline: 2.61×
- Verified on: Apple M5
- Sampler: temperature 0.6 · top_p 0.95 · top_k 20
See mtplx_runtime.json for the full verification record.
Usage
# MTPLX picks this model up automatically when downloaded:
mtplx pull FlatFootInternational/qwen-qwen3.5-9b-MTPLX-8bit-g32
mtplx start chat
License
See LICENSE.
- Downloads last month
- 40
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support