Updated daily at 05:00 UTC
Vision-language
GPT-5
- Released
- 7 Aug 2025
- Status
- Deprecated
- Retired
- 11 Dec 2026
- Context
- 400K
- Max output
- 128K
- Input per 1M tokens
- $1.25
- Output per 1M tokens
- $10.00
- Open weights
- No
- License
- proprietary
- Parameters
- —
- Knowledge cutoff
- 30 Sep 2024
- Reasoning
- Yes
- Modalities
- text, image → text
Category scores
Out of 100, from independent and third-party results only, by method category-2026-09-29.1.
| Category | Score | Rank | Coverage |
|---|---|---|---|
| Knowledge & reasoning | 35.4 | 41 | 4 of 4 |
| Coding & software engineering | 47.4 | Not ranked | 2 of 5 |
| Agentic tool use | 41.2 | 7 | 2 of 2 |
| Vision & multimodal understanding | 65.0 | 3 | 2 of 3 |
| Factuality & hallucination | 50.1 | 22 | 1 of 1 |
| Human preference | 44.2 | 39 | 3 of 5 |
Benchmark results
Every result we hold for this model, with who ran it and where it was published.
Knowledge & reasoning
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| Chess Puzzles | Puzzles solved | 37.0% | Independent | high | 8 Dec 2025 | Epoch AI |
| Chess Puzzles | Puzzles solved | 29.0% | Independent | medium | 15 Jul 2026 | Epoch AI |
| Chess Puzzles | Puzzles solved | 24.0% | Independent | low | 15 Jul 2026 | Epoch AI |
| Chess Puzzles | Puzzles solved | 16.0% | Independent | minimal | 15 Jul 2026 | Epoch AI |
| DTBench (Decision Theory Benchmark) | Accuracy (capability questions) | 90.7% | Third party | high | — | https://conceptualreasoning.ai/dtbench via Epoch AI |
| Earthborne Rangers (EBR-bench) | Objectives reached | 12.7% | Independent | high | 29 Jun 2026 | Epoch AI |
| LMCA (Language Model Conceptual Argumentation) | Correlation with expert ratings (Pearson r) | 0.3999 | Third party | high | — | https://conceptualreasoning.ai/lmca via Epoch AI |
| Mystery Game Puzzles | Puzzles solved | 23.0% | Independent | high | 5 Aug 2026 | Epoch AI |
| Mystery Game Puzzles | Puzzles solved | 16.0% | Independent | low | 27 Aug 2026 | Epoch AI |
| Mystery Game Puzzles | Puzzles solved | 15.0% | Independent | medium | 27 Aug 2026 | Epoch AI |
| Mystery Game Puzzles | Puzzles solved | 14.0% | Independent | minimal | 27 Aug 2026 | Epoch AI |
| SimpleBench | Score (AVG@5) | 56.7% | Third party | high | — | https://lmcouncil.ai/benchmarks via Epoch AI |
Scientific reasoning
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| GPQA Diamond | Accuracy | 86.2% | Independent | high | 29 Oct 2025 | Epoch AI |
| GPQA Diamond | Accuracy | 85.4% | Independent | medium | 7 Aug 2025 | Epoch AI |
| GPQA Diamond | Accuracy | 71.7% | Independent | minimal | 20 Jul 2026 | Epoch AI |
Mathematics
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| FrontierMath Tier 4 | Accuracy | 22.0% | Independent | high | 11 Jun 2026 | Epoch AI |
| FrontierMath Tier 4 (v1) | Accuracy | 12.5% | Independent | high | 30 Oct 2025 | Epoch AI |
| FrontierMath Tier 4 (v1) | Accuracy | 6.3% | Independent | medium | 7 Aug 2025 | Epoch AI |
| FrontierMath Tiers 1-3 | Accuracy | 55.4% | Independent | high | 10 Jun 2026 | Epoch AI |
| FrontierMath Tiers 1-3 | Accuracy | 37.2% | Independent | low | 27 Aug 2026 | Epoch AI |
| FrontierMath Tiers 1-3 | Accuracy | 18.2% | Independent | minimal | 27 Aug 2026 | Epoch AI |
| FrontierMath Tiers 1-3 (v1) | Accuracy | 32.4% | Independent | high | 13 Nov 2025 | Epoch AI |
| FrontierMath Tiers 1-3 (v1) | Accuracy | 27.2% | Independent | medium | 13 Nov 2025 | Epoch AI |
| MATH Level 5 | Accuracy | 98.1% | Independent | high | 29 Oct 2025 | Epoch AI |
| MATH Level 5 | Accuracy | 97.9% | Independent | medium | 20 Aug 2025 | Epoch AI |
| OTIS Mock AIME 2024-2025 | Accuracy | 91.4% | Independent | high | 29 Oct 2025 | Epoch AI |
| OTIS Mock AIME 2024-2025 | Accuracy | 87.2% | Independent | medium | 7 Aug 2025 | Epoch AI |
| OTIS Mock AIME 2024-2025 | Accuracy | 46.7% | Independent | minimal | 20 Jul 2026 | Epoch AI |
Coding & software engineering
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| Aider Polyglot | Percent correct | 88.0% | Third party | high | — | https://aider.chat/docs/leaderboards/ via Epoch AI |
| Aider Polyglot | Percent correct | 86.7% | Third party | medium | — | https://aider.chat/docs/leaderboards/ via Epoch AI |
| Aider Polyglot | Percent correct | 81.3% | Third party | low | — | https://aider.chat/docs/leaderboards/ via Epoch AI |
| GSO | OPT@1 | 6.9% | Third party | high | — | https://gso-bench.github.io/index.html via Epoch AI |
| METR Time Horizons | Average task success (Epoch's column) | 69.6% | Third party | medium | — | METR - Measuring AI Ability to Complete Long Tasks via Epoch AI |
| METR Time Horizons | Average task success (Epoch's column) | 69.4% | Source not stated | high | — | Epoch AI via Epoch AI |
| SWE-bench Verified | % resolved | 73.6% | Independent | high | 6 Feb 2026 | Epoch AI |
| SWE-bench Verified | % resolved | 71.5% | Independent | medium | 5 Feb 2026 | Epoch AI |
| Terminal-Bench 2.0 | Accuracy (mean over trials) | 49.6% | Third party | unknown | — | https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI |
| Terminal-Bench 2.0 | Accuracy (mean over trials) | 49.6% | Third party | medium | — | Terminal-Bench v2 Leaderboard via Epoch AI |
| Terminal-Bench 2.0 | Accuracy (mean over trials) | 43.8% | Third party | unknown | — | https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI |
| Terminal-Bench 2.0 | Accuracy (mean over trials) | 43.8% | Third party | medium | — | Terminal-Bench v2 Leaderboard via Epoch AI |
| Terminal-Bench 2.0 | Accuracy (mean over trials) | 35.2% | Third party | medium | — | Terminal-Bench v2 Leaderboard via Epoch AI |
| Terminal-Bench 2.0 | Accuracy (mean over trials) | 35.2% | Third party | unknown | — | https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI |
| Terminal-Bench 2.0 | Accuracy (mean over trials) | 33.9% | Third party | unknown | — | https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI |
| Terminal-Bench 2.0 | Accuracy (mean over trials) | 33.9% | Third party | medium | — | Terminal-Bench v2 Leaderboard via Epoch AI |
| WeirdML (v2) | Average accuracy | 60.7% | Third party | high | — | https://htihle.github.io/weirdml.html via Epoch AI |
Agentic tool use
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| BALROG | Average progress | 32.8% | Third party | minimal | — | https://balrogai.com/ via Epoch AI |
| Deep Research Bench (FutureSearch) | Average score | 49.6% | Third party | low | — | https://evals.futuresearch.ai/#drb via Epoch AI |
| Deep Research Bench (FutureSearch) | Average score | 48.9% | Third party | minimal | — | https://evals.futuresearch.ai/#drb via Epoch AI |
| Deep Research Bench (FutureSearch) | Average score | 48.6% | Third party | medium | — | https://evals.futuresearch.ai/#drb via Epoch AI |
| Deep Research Bench (FutureSearch) | Average score | 48.1% | Third party | high | — | https://evals.futuresearch.ai/#drb via Epoch AI |
Long context
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| Fiction.LiveBench | % correct at 16K tokens (Epoch's column) | 97.2% | Third party | medium | — | Fiction.live leaderboard via Epoch AI |
Vision & multimodal understanding
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| GeoBench | Country accuracy | 81.0% | Third party | medium | — | GeoBench leaderboard via Epoch AI |
| Visual Physics Comprehension Test (VPCT) | Correct | 66.0% | Third party | high | — | VPCT leaderboard via Epoch AI |
| Visual Physics Comprehension Test (VPCT) | Correct | 63.2% | Third party | medium | — | VPCT leaderboard via Epoch AI |
Factuality & hallucination
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| SimpleQA Verified | Accuracy (% correct) | 50.1% | Independent | high | 27 Aug 2026 | Epoch AI |
Writing
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| LLM Creative Story-Writing Benchmark V1 (Lech Mazur) | Mean rubric score (percent of the 10-point scale) | 86.0% | Third party | medium | — | lechmazur/writing Github repository via Epoch AI |
Human preference
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| Code Arena: WebDev (arena.ai) | Arena score (Elo scale) | 1418 | Independent | medium | 25 Sep 2026 | Arena |
| Code Arena: WebDev (arena.ai) | Votes | 3090 | Independent | medium | 25 Sep 2026 | Arena |
| Text Arena (arena.ai) | Arena score (Elo scale) | 1435 | Independent | high | 25 Sep 2026 | Arena |
| Text Arena (arena.ai) | Votes | 31503 | Independent | high | 25 Sep 2026 | Arena |
| Vision Arena (arena.ai) | Arena score (Elo scale) | 1210 | Independent | high | 28 Sep 2026 | Arena |
| Vision Arena (arena.ai) | Votes | 34464 | Independent | high | 28 Sep 2026 | Arena |
Other GPT-5 versions
- GPT-5.6 Cyber10 Aug 2026
- GPT-5.6 Luna9 Jul 2026
- GPT-5.6 Sol9 Jul 2026
- GPT-5.6 Terra9 Jul 2026
- GPT-5.523 Apr 2026
- GPT-5.5 Pro23 Apr 2026
- GPT-5.4 mini17 Mar 2026
- GPT-5.4 nano17 Mar 2026
- GPT-5.45 Mar 2026
- GPT-5.4 Pro5 Mar 2026
- GPT-5.3 Instant (gpt-5.3-chat-latest)3 Mar 2026
- GPT-5.211 Dec 2025
- GPT-5.2 Instant (gpt-5.2-chat-latest)11 Dec 2025
- GPT-5.2 Pro11 Dec 2025
- GPT-5.112 Nov 2025
- GPT-5.1 Instant (gpt-5.1-chat-latest)12 Nov 2025
- GPT-5 Instant (gpt-5-chat-latest)7 Aug 2025
- GPT-5 mini7 Aug 2025
- GPT-5 nano7 Aug 2025
- GPT-5 Pro7 Aug 2025
- GPT-5.4 Cyber—
- GPT-5.5 Cyber—