Single AMD R9700 Hits 153 tok/s Running Qwen3.8 27B NVFP4 After Optimization
whodoneit1 · reddit · 2026-09-17
A Reddit user doubled inference performance for a single AMD Radeon R9700 card running Unsloth's Qwen3.8 27B NVFP4: peak decode of 153 tok/s (JSON tasks), 470 tok/s at 8 concurrent requests, and 3,619 tok/s prefill.
Per-category results:
- : 153.1 tok/s (fastest); summarization: 141.7; math: 140; fileedit: 138
- reasoning: 123.9; code: 120.5
- chat: 67.1; prose: 69.2
p50 update latency held steady at 34-42ms across tasks. Results and charts are published in the BetterBench GitHub project — a useful reference for single-GPU AMD setups.
Related event: Single AMD R9700 Hits 153 tok/s on Qwen3.8 27B with NVFP4(2 posts)→
More from Infra
- To dethrone Postgres, new databases must prioritize ecosystem integrations from day one — mgill25 · 2026-11-02
- ESP32-powered 4-inch touchscreen builds a TUI dashboard for monitoring a DGX inference cluster — CATLLM · 2026-11-02
- OpenRouter crosses 10 quadrillion annualized token run rate, doubling every 10.5 weeks — deedydas · 2026-11-02
- Palmer Luckey asks: can you use datacenter RAM for things prohibited by the operator's terms of service? — avaitopiper · 2026-11-02
- Disable MTP for 20-30% faster local LLM inference on limited RAM — InternalMode8159 · 2026-11-02
- Kioxia up 216% in a year but down 68% from its June 2026 peak — zephyr_z9 · 2026-11-02