Skip to content
llm-speed

Qwen2.5-Coder-32B-Instruct vs Codestral-22B-v0.1

Signed, community-submitted decode tok/s, prefill, and TTFT for Qwen2.5-Coder-32B-Instruct vs Codestral-22B-v0.1 — every number links to the run it came from.

Verdict

On the RTX 5090 (32GB), Codestral-22B-v0.1 decodes at 100 tok/s versus 72 tok/s for Qwen2.5-Coder-32B-Instruct, 1.4× faster. Across 2 hardware configs measured on both, Codestral-22B-v0.1 is faster on 2 of 2. Every cell in the table below links to the submitted run it came from.

model
Qwen2.5-Coder-32B-Instruct
Qwen · 32B
View Qwen2.5-Coder-32B-Instruct page →
model
Codestral-22B-v0.1
Mistral · 22B
View Codestral-22B-v0.1 page →

Hardware with data on both models

HardwareQwen2.5-Coder-32B-Instruct decodeCodestral-22B-v0.1 decodeΔSource runs
RTX 5090 (32GB)71.91tok/s100.3tok/s-28.4r_nkbs6d3-d21 · r_4q040m4scic
M3 Ultra (60-core GPU)34.48tok/s47.49tok/s-13.0r_721b4bls_oq · r_79dvtag5fd_

See also: Qwen2.5-Coder-32B-Instruct benchmarks · Codestral-22B-v0.1 benchmarks · All hardware · All models · Methodology