Last released Apr 15, 2026
39% faster TTFT, 67% less KV cache, zero config — autotune optimises local LLMs on Ollama, LM Studio, and MLX
Supported by