GPT-5.1 vs GPT-5.2
Respan Index by category
Each category is scored 0 to 100, and how far ahead the winner is. Leads under a point are close to even.
| Category | GPT-5.1 | GPT-5.2 |
|---|---|---|
| Coding | 14.7 GPT-5.1 | 26.0 GPT-5.2 |
| Agents | 33.9 GPT-5.1 | 59.5 GPT-5.2 |
| Reasoning | 40.2 GPT-5.1 | 55.3 GPT-5.2 |
| Knowledge & language | 47.9 GPT-5.1 | 38.1 GPT-5.2 |
| Human preference | 51.1 GPT-5.1 | 51.5 GPT-5.2 |
Benchmarks both report
GPT-5.2 leads on 11 of the 15 benchmarks both report. Higher is better. Each score links to where it was published.
| Benchmark | GPT-5.1 | GPT-5.2 |
|---|---|---|
| LMArena Elo | 1456 | 1437 |
| LMArena Vision (LMArena) | 1251 | 1243 |
| LMArena WebDev (LMArena) | 1397 | 1419 |
| Chess Puzzles (Epoch AI) | 42% | 48% |
| GPQA Diamond | 87.6% | 92.4% |
| Mystery Game Puzzles (Epoch AI) | 19% | 23% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 88.6% | 96.1% |
| SAGE (Vals AI) | 43.2% | 49.3% |
| SimpleQA Verified | 48% | 37.1% |
| SWE-bench Verified (Epoch AI) | 68% | 73.8% |
| Vending-Bench 2 (Andon Labs) | 1473.43 | 3591.33 |
| WeirdML (Håvard Tveit Ihle) | 60.8% | 72.2% |
| AIME 2025 | 94.2% | 100% |
| GSO Opt@1 (GSO) | 12.8% | 26.5% |
| SimpleBench (SimpleBench) | 53.2% | 45.8% |
FAQ
Is GPT-5.1 better than GPT-5.2?
GPT-5.2 scores 46.7 on the Respan Index (#51), ahead of GPT-5.1 at 34.7 (#81). GPT-5.2 leads on 11 of the 15 benchmarks both report.
Which is cheaper, GPT-5.1 or GPT-5.2?
GPT-5.1 lists at $1.25 input and $10 output per 1M tokens; GPT-5.2 at $1.75 and $14. GPT-5.1 is 1.4x cheaper per token at list price.
Which has the larger context window?
Both accept 400K tokens of context.