ComfyUI v0.45.0 speeds up local Qwen3.x text generation by up to two times
Understand this story
In v0.45.0, ComfyUI rewrote the inference hot path for Qwen3.x text generation, delivering up to a 2x increase in token generation speed on the same hardware and with the same weights. The figures given in the ComfyUI blog are: on an RTX 5090, Qwen3.5 9B goes from 106 to 219 tokens/s, and on an RTX 5080 from 91 to 129 tokens/s; Qwen3.8 27B goes from 55 to 122 tokens/s when generating a 1536×1536 image description. In practice, load a Qwen3.x checkpoint and keep the Text Generate node's mtp option set to auto; you can also launch with --use-ck-attention to enable Comfy-Kitchen attention. ComfyUI says these improvements cover NVIDIA and AMD, Windows and Linux, and require no extra steps beyond updating to the latest version.
AI overview based on source reports · Updated 3 days ago
Coverage timeline
Follow the coverage to explore different perspectives.
- Comfy BlogDoubling LLM token rates in ComfyUI
ComfyUI rewrote the hot path of local text generation for Qwen3.x LLMs, letting the same checkpoints on the same GPU generate tokens up to twice as fast with no change to model or output quality; on an RTX 5090, Qwen3.5 9B went from 106 to 219 tokens per second. The changes include a re-laid-out GEMV kernel, fused prologue operations, new attention and DeltaNet decode kernels, and an adaptive-tree MTP speculative decoding algorithm, all available in ComfyUI v0.45.0 with the Text Generate node's mtp option left on auto.
Activity for this story
- Current comparable activity
- 3
- Peak comparable activity
- 10Nov 12 02:00
- Change in the past 24 hours
- -49%
Trends compare the same fully observed participants, which may cover fewer participants than current activity. Move the pointer or tap the chart to inspect hourly activity; use the left and right arrow keys with a keyboard.