Qwen3.6 NVFP4 Gets 2.5x Speed Boost
danielhanchen · x · 2026-07-10
Unsloth has leveraged its dynamic approach to achieve faster inference speeds for Qwen3.6 using NVFP4 quantization.
The results shared include:
- The NVFP4 version of Qwen3.6-27B is 2.5x faster than the original implementation
- Comparing W4A4 to W4A16
- Two versions for the 35B-A3B model:
- NVFP4: 1.56x faster
- NVFP4-Fast: 1.79x faster
The Fast version uses pure W4A4, while the non-Fast version mixes different precisions to strike a balance between speed and accuracy. The original post also notes that these quantized versions improve accuracy, tool calling, agent use, and looping performance.
Related event: Unsloth Brings Faster NVFP4 Quantization to Qwen3.6(3 posts)→
More from Infra
- LLM Serving Metrics Thread: Why TPOT and Uptime Make or Break User Experience — abhijithneil · 2026-09-11
- PlanetScale launches sharded Postgres: 768 servers acting as one, 1PB scale — dhruv2038 · 2026-09-11
- Can a 7900 XTX 24GB run Qwen locally? Reddit seeks ROCm tok/s benchmarks — thenomadexplorerlife · 2026-09-11
- RTK Terminal Compression Cuts Tokens but Leaves Your AI Coding Bill Unchanged — Bartaseth · 2026-09-11
- SF Compute founder: buying compute is 'an absolutely awful experience' right now — IgorCarron · 2026-09-11
- SmolVM open-sources persistent computer infrastructure for agents that outlive chat sessions — aniketmaurya · 2026-09-11