GPT-6.2 Astra
Complex projects. Difficult software, computer-use and document workflows.
Which reasoning effort setting?
controls how long GPT-6.2 Astra thinks before answering. Token prices never change; higher settings spend more and raise task cost.
Start at High. It scores 60.1 against 62.1 at the top setting, for 48% less per benchmark task. Rule: cheapest setting within 2 points of the model’s best score.
At High, GPT-6.2 Astra thinks for about 7.9K tokens per task, then writes about 9.9K more as its answer: $2.07 per task, 3.1× Low, for +14.4 points.
Output rate: $40.00/1M at every effort level. Thinking tokens are billed as output.
| Effort | Index | Per task | Thinking | All output | × Low | Time |
|---|---|---|---|---|---|---|
| Low | 45.7 | $0.66 | 1.5K | 5.2K | 1.00× | 2 min |
| Medium | 57.0 | $1.62 | 5.5K | 13.2K | 2.44× | 4.9 min |
| HighStart here | 60.1 | $2.07 | 7.9K | 17.8K | 3.13× | 6.4 min |
| Xhigh | 61.6 | $2.52 | 11.1K | 22.2K | 3.81× | 7.8 min |
| Max | 62.1 | $4.00 | 26.1K | 41K | 6.04× | 14.2 min |
Per task of the Artificial Analysis Intelligence Index: the average over its nine evaluations, counting input, cache, and output tokens at list prices. These are long, agent-style benchmark tasks, much longer than a chat message.
Where GPT-6.2 Astra sits on the frontier
Intelligence Index against one 100K-input, 10K-output request.