Skip to content

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

llm: Enable flash attention for mistral3 by default

Jesse Gross9191dfaf05ea
diff --git a/fs/ggml/ggml.go b/fs/ggml/ggml.go
index 6ce9724f2..4004bbfd9 100644
--- a/fs/ggml/ggml.go
+++ b/fs/ggml/ggml.go
@@ -831,6 +831,7 @@ func (f GGML) FlashAttention() bool {
return slices.Contains([]string{
"gemma3",
"gptoss", "gpt-oss",
+ "mistral3",
"qwen3", "qwen3moe",
"qwen3vl", "qwen3vlmoe",
}, f.KV().String("general.architecture"))