The fastest Qwen model is Qwen3 32B at 328 output tokens per second. Qwen3.6 35B A3B (172 t/s) and Qwen3 Next 80B A3B Thinking (172 t/s) round out the top three.
AI models ranked by output speed (tokens per second, p50). The fastest large language models for low-latency and high-throughput applications.
The fastest Qwen model is Qwen3 32B at 328 output tokens per second. Qwen3.6 35B A3B (172 t/s) and Qwen3 Next 80B A3B Thinking (172 t/s) round out the top three.
Qwen3.6 35B A3B (172 t/s) is the closest alternative on this metric, followed by Qwen3 Next 80B A3B Thinking (172 t/s). See the full ranking above for the tradeoffs.
modelgrep tracks 49 Qwen models with live benchmarks, speed, latency and per-provider pricing, led on intelligence by Qwen3.7 Max. 25 of them qualify for this ranking.