Claude Sonnet 4.6 (Non-reasoning, High Effort) vs GPT-5.2 (xhigh)
For product management work, compared on the PM Index.
- GPT-5.2 (xhigh) leads on the PM Index (52.8 vs 48.0).
- GPT-5.2 (xhigh) is cheaper ($4.81 vs $6.00 per 1M tokens).
- GPT-5.2 (xhigh) is faster (78 vs 43 tok/s).
| Metric | Claude Sonnet 4.6 (Non-reasoning, High Effort) | GPT-5.2 (xhigh) |
|---|---|---|
| PM Index | 48.0 | 52.8 |
| AA Intelligence | 44.4 | 51.3 |
| AA Coding | 46.4 | 48.7 |
| AA Agentic | 61.6 | 60.2 |
| Blended $/1M | $6.00 | $4.81 |
| Output tok/s | 43 | 78 |