Codestral 25.08 vs GPT-5.6 Luna
GPT-5.6 Luna reads a 26.2K-token pull request for $0.0052, 33% less than Codestral 25.08 ($0.0079). Above 272K tokens long-context rates apply, and Codestral 25.08 becomes the cheaper one. Both sit in the economy price tier, so the choice between them is about review quality more than budget.
Side by side
| Figure | Codestral 25.08 | GPT-5.6 Luna |
|---|---|---|
| Price tier | Economy | Economy |
| Typical review calllist price | $0.01 | $0.0081 (better) |
| Input / 1M tokens | $0.30 | $0.20 (better) |
| Output / 1M tokens | $0.90 (better) | $1.20 |
| Context window | 128K | 1.1M |
- Above 272K tokens, long-context rates mean Codestral 25.08 is the cheaper of the two.
Not benchmarked yet
Codestral 25.08 and GPT-5.6 Luna aren't benchmarked yet, so this page compares list prices only. Measured bugs caught, false alarms and a bug-by-bug comparison appear here once both have results.
See measured models