Unsloth makes local inference with Qwen3.8-Flash 1.7 times faster
Qwen3.8-Flash can now run 1.7× faster locally with MTP!⚡️ GGUFs can reach 170 tokens/s on a RTX PRO...
AI BriefingUsing MTP (Multi-Token Prediction), Unsloth makes local inference with Qwen3.8-Flash-Next approximately 1.3 to 1.7 times faster without changing accuracy. GGUF can reach 170 tokens/s on a single RTX PRO 6000, compared with a baseline of 100 tokens/s.
Source: Unsloth · x.com
Event follow-up1
View all event coverage- Unsloth makes local inference with Qwen3.8-Flash 1.7 times fasterUnsloth · September 2