Qwen3-4B-Instruct-2507-4bit
4 workload results across 1 hardware configuration.
Fastest community-submitted config (not measured by llm-speed)
182.7 decode tok/s
on M3 Ultra (60-core GPU) + 96GB unified via mlx. see full run
Signed runs (4 workload results)
This run is community-submitted (signed on the submitter's own machine, not measured by llm-speed, and not used for llm-speed verdicts).
M3 Ultra (60-core GPU) + 96GB unified
| Workload | Backend | Quant | decode tok/s | prefill tok/s | TTFT | Run | Source |
|---|---|---|---|---|---|---|---|
| chat-short | mlx@0.31.3 | 4BIT | 182.7tok/s | 636.6tok/s | 173ms | r_p2rpb_0iyjj | Community-submitted |
| chat-long | mlx@0.31.3 | 4BIT | 147.0tok/s | 1,913.7tok/s | 1,644ms | r_p2rpb_0iyjj | Community-submitted |
| concurrent-decode | mlx@0.31.3 | 4BIT | 168.4tok/s | no data | no data | r_p2rpb_0iyjj | Community-submitted |
| agent-trace | mlx@0.31.3 | 4BIT | 160.8tok/s | 1,913.6tok/s | 1,091ms | r_p2rpb_0iyjj | Community-submitted |
Can Qwen3-4B-Instruct-2507-4bit run on your hardware?
Fastest decode speed on each GPU or Mac, from llm-speed's own runs and community-submitted runs. Each page lists every quant and runtime and labels which runs llm-speed measured.
- M3 Ultra · 182.7 tok/s
Qwen3-4B-Instruct-2507-4bit on hardware
Check it for your setup
- Predict your exact setup: estimate whether Qwen3-4B-Instruct-2507-4bit fits your GPU or Mac at your quant and context, next to measured speeds where we have them.
- Request this pair: ask for a measured Qwen3-4B-Instruct-2507-4bit run on hardware we have not measured. Requests are ranked by votes.
- Full run table (1 run): every Qwen3-4B-Instruct-2507-4bit result we have, with links to each source run.