Get up and running with Kimi, GLM, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.
Commits on main
mlx: respect tokenizer add_bos_token setting in pipeline (#15185)
4d14b0f Daniel Hiltgen committed 7 months ago
model: add qwen3-next compatibility for legacy ssm_in projections (#15133)
b7bda92 Jeffrey Morgan committed 7 months ago
launch: improve multi-select for already added models (#15113)
7c8da56 Parth Sareen committed 7 months ago
launch: auto-install pi and manage web-search lifecycle (#15118)
6214103 Parth Sareen committed 7 months ago
server: preserve raw manifest bytes during pull (#15104)
3824e38 Bruce MacDonald committed 7 months ago
launch: skip context length warning for MLX models and show model name (#15102)
b00bd1d Parth Sareen committed 7 months ago
anthropic: fix KV cache reuse degraded by tool call argument reordering
ac83ac2 Jesse Gross committed 8 months ago
parsers: qwen3.5 streaming tool-call parsing and add regression test (#15098)
69ed0c2 Jeffrey Morgan committed 7 months ago
model/parsers: close think block if tool block starts in Qwen3.5 (#15022)
1cefa74 Alfredo Matas committed 7 months ago
launch: warn when server context length is below 64k for local models (#15044)
366625a Eva H committed 7 months ago
launch/vscode: prefer known vs code paths over `code` on PATH (#15073)
1adfc27 Eva H committed 7 months ago
mlxrunner: combine setStateRaw and setStateDetached into setState
9d7b18f Jesse Gross committed 8 months ago