OpenAI·released 2025-10-071 source
GPT-5 Pro benchmark scores: 8 benchmarks tracked. 25% of its scores are independently reproduced — the rest are self-reported or unverified.
Leads on 0 of 8 benchmarks·best result 70.2 on ARC-AGI (unverified)·2 of 8 independently reproduced·$15/$120 per M tokens
Consensus: LiteLLM · Cross-check: models.dev, OpenRouter · See every model’s pricing →
| Benchmark | Score | Evidence status | Measured |
|---|---|---|---|
| ARC-AGI | 70.2accuracy (%) | unverified· optimizedT1 | 2025-10-07ARC Prize Leaderboard |
| WeirdML | 60.39accuracy (%) | unverifiedT1 | 2025-10-07https://htihle.github.io/weirdml.html |
| FrontierMath-Tiers-1-3-v2-Private | 55.79accuracy (%) | reproducedT1 | 2025-10-07Epoch AI |
| SimpleBench | 53.92accuracy (%) | unverifiedT1 | 2025-10-07SimpleBench Leaderboard |
| Chess Puzzles | 45.29accuracy (%) | unverifiedT2 | 2025-10-07 |
| HLE | 28.19accuracy (%) | unverifiedT2 | 2025-10-07 |
| FrontierMath-Tier-4-v2-Private | 19.51accuracy (%) | reproducedT1 | 2025-10-07Epoch AI |
| ARC-AGI-2 | 18.33accuracy (%) | unverifiedT2 | 2025-10-07 |
Claims drawn from cited facts, not live model generation.
This model was developed by OpenAI and released in 2025.
2 cited facts
This model has scores on 8 tracked benchmarks and currently leads 0 of them. The record’s evidence status is independently reproduced.
3 cited facts
Its mean gap to the best recorded score across its tracked benchmarks is 41.57 points.
1 cited fact
6 facts cross-checked across data sources: 3 corroborated, 2 single-source, 1 disputed.
Source facts, citations, and refresh stamp for this record.
Sources: epoch_benchmarks, litellm_prices, modelsdev_models, openrouter_models
GPT-5 Pro is an AI model developed by OpenAI, released 2025-10-07. Its benchmark record below tags every score as independently reproduced, self-reported, or unverified.
GPT-5 Pro has recorded scores on 8 benchmarks, each shown with its evidence status.
GPT-5 Pro has recorded scores on 8 benchmarks — WeirdML, Chess Puzzles, SimpleBench, ARC-AGI, ARC-AGI-2, HLE, and 2 more. The full table above shows each score with its evidence status.
2 of 8 recorded scores (25%) are independently reproduced rather than self-reported by the lab.
GPT-5 Pro has 8 tracked claims: 2 independently reproduced, 6 unverified.
Listed API pricing: $15 per million input tokens, $120 per million output tokens. See the pricing block for the full breakdown.