Codestral-22B-v0.1
5 workload results across 2 hardware configurations.
Model weights used in these runs:
- mlx-community/Codestral-22B-v0.1-4bitView on Hugging Face(1 of 2 runs)
Fastest community-submitted config (not measured by llm-speed)
93.7decode tok/s
on RTX 5090 (32GB) via llama.cpp. see full run
Can Codestral-22B-v0.1 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.
Signed runs (5 workload results)
All 2 runs are community-submitted (signed on the submitter's own machine, not measured by llm-speed, and not used for llm-speed verdicts).
4 results · fastest chat-short decode 93.7 tok/s
| Workload | Backend | Quant | decode tok/s | prefill tok/s | TTFT | Run | Source |
|---|---|---|---|---|---|---|---|
| chat-short | llama.cpp | - | 93.72tok/s | no data | 68.6ms | r_4q040m4scic | Community-submitted |
| chat-long | llama.cpp | - | 87.36tok/s | no data | 849ms | r_4q040m4scic | Community-submitted |
| concurrent-decode | llama.cpp | - | 100.3tok/s | no data | no data | r_4q040m4scic | Community-submitted |
| agent-trace | llama.cpp | - | 93.44tok/s | 10,351.0tok/s | 191ms | r_4q040m4scic | Community-submitted |
1 result · fastest chat-short decode 47.5 tok/s
| Workload | Backend | Quant | decode tok/s | prefill tok/s | TTFT | Run | Source |
|---|---|---|---|---|---|---|---|
| chat-short | mlx@0.31.3 | 4BIT | 47.49tok/s | 205.7tok/s | 559ms | r_79dvtag5fd_ | Community-submitted |
Codestral-22B-v0.1 on hardware
Request a Codestral-22B-v0.1 run
No measured run yet for these pairs. Vote for the one you want; the most-requested runs go first. See all requests.
0 votes
We never sell or share your email. Privacy
llm-speed can measure this card · Submit your own
Compare Codestral-22B-v0.1 head-to-head
Check it for your setup
- Predict your exact setup: estimate whether Codestral-22B-v0.1 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 Codestral-22B-v0.1 run on hardware we have not measured. Requests are ranked by votes.
- Full run table (2 runs): every Codestral-22B-v0.1 result we have, with links to each source run.