DeepSeek V3.2 Exp (Non-reasoning) vs Mistral Large 3
For product management work, compared on the PM Index.
- DeepSeek V3.2 Exp (Non-reasoning) leads on the PM Index (29.1 vs 22.6).
- DeepSeek V3.2 Exp (Non-reasoning) is cheaper ($0.31 vs $0.75 per 1M tokens).
| Metric | DeepSeek V3.2 Exp (Non-reasoning) | Mistral Large 3 |
|---|---|---|
| PM Index | 29.1 | 22.6 |
| AA Intelligence | 28.4 | 22.8 |
| AA Coding | 30.0 | 22.7 |
| AA Agentic | 31.0 | 21.7 |
| Blended $/1M | $0.31 | $0.75 |
| Output tok/s | — | 56 |