Codestral 2508 (batch) vs GPT-5.6 Sol
Side-by-side API pricing and Artificial Analysis performance for Codestral 2508 (batch) (Mistral) and GPT-5.6 Sol (OpenAI). Green cells mark the better value in each row.
Prices updated Sep 14, 2026, 6:47 AM UTC · refreshed hourly
| Codestral 2508 (batch) Mistral | GPT-5.6 Sol OpenAI | |
|---|---|---|
| Pricing | ||
| Input $/1M | $0.15 | $2.00 |
| Output $/1M | $0.45 | $10.00 |
| Blended $/1M (3:1) | $0.225 | $4.00 |
| RAG example (30K in / 2K out) | $0.0054 | $0.08 |
| Quality & speed | ||
| Intelligence Index | — | 47.1 |
| Coding Index | — | 77.4 |
| Agentic Index | — | 50.5 |
| Output speed (tok/s) | — | 68 |
| Time to first token | — | 73.86s |
| Specs | ||
| Context window | 256K | 1.05M |
| Input modalities | text, file | file, image, text |
FAQ
Which is cheaper, Codestral 2508 (batch) or GPT-5.6 Sol?
Codestral 2508 (batch) has the lower blended API price at $0.225 per 1M tokens (3:1 input:output mix), versus $4.00 for the other model. Prices are live from OpenRouter and refresh about hourly.
Which is better for RAG workloads?
For a typical RAG request (30K input / 2K output tokens), Codestral 2508 (batch) costs about $0.0054 per request versus $0.08. Also compare context windows: Codestral 2508 (batch) offers 256K and GPT-5.6 Sol offers 1.05M.
Which scores higher on Artificial Analysis benchmarks?
GPT-5.6 Sol leads on the Artificial Analysis Intelligence Index (47.1 vs —). Check coding and agentic indices on this page for workload-specific tradeoffs.