Codestral 2508 is 2.9× cheaper per million tokens (blended). Qwen3.8 27B has the larger context window (1,000,000 vs 256,000 tokens, 3.9×). Qwen3.8 27B is open-weights; the other is not.

Codestral 2508 vs Qwen3.8 27B

Live catalog fields. Best-in-row is highlighted.

Codestral 2508

Mistral · mistralai__codestral-2508

Qwen3.8 27Bopen

Qwen · qwen__qwen3.8-27b

Identity

Fieldmistralai__codestral-2508qwen__qwen3.8-27b
ProviderMistralQwen
Slugmistralai__codestral-2508qwen__qwen3.8-27b
Statuslivelive
Open weightsNoYes
License
HuggingFaceQwen/Qwen3.8-27B

Quality

Fieldmistralai__codestral-2508qwen__qwen3.8-27b
Quality index (0–100)87.4

Cost

Fieldmistralai__codestral-2508qwen__qwen3.8-27b
Input $ / 1M tokens$0.30$0.42
Cached input $ / 1M$0.03$0.09
Output $ / 1M tokens$0.90$3.00

Context

Fieldmistralai__codestral-2508qwen__qwen3.8-27b
Max context tokens256,0001,000,000
Max output tokens204,800131,072

Modalities

Fieldmistralai__codestral-2508qwen__qwen3.8-27b
Inputtext, filetext, image, video
Outputtexttext