GPT-5.2 (Thinking): benchmark results
Provider: OpenAI. Released 2025-12-11. Access: API.
Aggregate Rating 1751 ± 10, rank #239 of 2170 rated models, from 133 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AgentNoiseBench (VitaBench, User Noise) | 45 | Success on a stratified 25 percent sample of VitaBench tasks | 100 |
| AgentNoiseBench (tau2-bench, Tool Noise) | 61 | Success on a stratified 25 percent sample of tau2-bench task | 100 |
| BenchPreS | 46.38 | Appropriate Application Rate minus Misapplication Rate, in p | 100 |
| BrowseComp-V3 (Official Web Platforms) | 39.13 | Success rate (%): share of questions answered correctly on 3 | 100 |
| BrowseComp-V3 (Official Web Platforms) - Process Score | 66.05 | Process score (%): mean share of each question's expert-anno | 100 |
| CCTU (Thinking) | 18.17 | Perfect solve rate (%): the share of cases in which the mode | 100 |
| Doc-PP - Explicit Query Faithfulness (No Policy) | 98 | Faithfulness (%; responses stating the requested value, expl | 100 |
| Doc-PP - Explicit Query Leakage | 8.1 | Leakage Rate (%; responses revealing the protected value, ex | 100 |
| Doc-PP - Implicit Query Leakage | 25.7 | Leakage Rate (%; responses revealing the protected value, im | 100 |
| Doc-PP - OCR Input - Explicit Query Leakage | 11.5 | Leakage Rate (%; responses revealing the protected value, ex | 100 |
| Doc-PP - OCR Input - Implicit Query Faithfulness (With Policy) | 70.6 | Faithfulness (%; content-checklist criteria met, implicit qu | 100 |
| Doc-PP - OCR Input - Implicit Query Leakage | 25.8 | Leakage Rate (%; responses revealing the protected value, im | 100 |
Interactive version: theaggregate.ai/model?slug=gpt-5-2-thinking · How It Works · Data refreshed daily, snapshot 2026-11-05.