Benchmarks
351 sources · 1,261 models × 3,053 benchmarks · 160,189 scoresThe LifeVal chart of OpenAI's GPT-6.2 launch post, which OpenAI calls LifeVal and links to its LifeVal page. Its values are not the LifeVal page's: GPT-6 Astra scores 67.5 to 79.8 across efforts here, against 66.6 to 82.5 there, at about half the cost per task, so the task set, grading or harness changed and is not stated. Kept apart from LifeVal until OpenAI documents the version. Diagnostic; OpenAI runs only.
No signal or difficulty estimate. Held out of the fit by design.
| # | Model | Setting | Score | Source |
|---|---|---|---|---|
| 1 | GPT-6.2 Astra | max | 86.1 | |
| 2 | GPT-6.2 Sol | max | 83.8 | |
| 3 | GPT-6 Astra | max | 79.8 | |
| 4 | GPT-6.1 Sol | max | 79.5 |