Skip to content

Get up and running with Kimi, GLM, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.

Commits on main

qwen3.8: add renderer and MLX import support (#17745)
Daniel Hiltgen committed 3 months ago
5512797
launch: add DeepSeek Harness integration (#17733)
Parth Sareen committed 3 months ago
39df91c
7ce88bd
launch: add Muse Code integration (#17594)
Daniel Hiltgen committed 3 months ago
01d04d5
9a56a0e
8831349
nn: speed up prefill on double-scale nvfp4 models
Jesse Gross committed 3 months ago
2b4a993
llama.cpp bump (#17702)
Daniel Hiltgen committed 3 months ago
e922bc7
MLX update (#17704)
Daniel Hiltgen committed 3 months ago
950dd9a
openai: support web search in Responses API (#17686)
Parth Sareen committed 3 months ago
b6b1b25
mlx: enable CUDA backend in CUDA builds (#17688)
Daniel Hiltgen committed 3 months ago
641df5e
docs: fix broken links (#17676)
Eva H committed 3 months ago
948f693
96fb6d2
mlx: implement Nemotron 3 Nano Omni (#17060)
Daniel Hiltgen committed 3 months ago
bb7bba8
llama.cpp update (#17659)
Daniel Hiltgen committed 3 months ago
4f066a6
docs: require VS Code 1.127 (#17655)
Eva H committed 3 months ago
a836eb8
1a9e423
Release v0.32.7 (#17646)
Daniel Hiltgen committed 3 months ago
43f4eda
MLX: version bump (#17637)
Daniel Hiltgen committed 3 months ago
acdf815
qwen3_5: image input support
Jesse Gross committed 3 months ago
1e85fe8
5fcf71b
mlxrunner: expand image tags into placeholder tokens
Jesse Gross committed 3 months ago
60bdc23
mlxrunner: fold media identity into prefix-trie keys
Jesse Gross committed 3 months ago
694487c
af5b627
8713570
5a173ed
laguna: wire the DFlash target side
Jesse Gross committed 3 months ago
b880b76