Updated daily at 05:00 UTC
Vision-language
Claude Opus 4.8
- Released
- 28 May 2026
- Status
- Legacy
- Context
- 1M
- Max output
- 128K
- Input per 1M tokens
- $5.00
- Output per 1M tokens
- $25.00
- Open weights
- No
- License
- proprietary
- Parameters
- —
- Knowledge cutoff
- Jan 2026
- 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 | 46.1 | 25 | 4 of 4 |
| Coding & software engineering | 50.8 | 3 | 3 of 5 |
| Agentic tool use | 50.2 | Not ranked | 1 of 2 |
| Computer & browser use | 20.6 | 3 | 1 of 1 |
| Vision & multimodal understanding | 17.9 | Not ranked | 1 of 3 |
| Factuality & hallucination | 53.0 | 17 | 1 of 1 |
| Human preference | 50.4 | 29 | 5 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 | 34.0% | Independent | max | 29 May 2026 | Epoch AI |
| Chess Puzzles | Puzzles solved | 29.0% | Independent | low | 6 Aug 2026 | Epoch AI |
| Chess Puzzles | Puzzles solved | 13.0% | Independent | none | 6 Aug 2026 | Epoch AI |
| DTBench (Decision Theory Benchmark) | Accuracy (capability questions) | 94.9% | Third party | max | — | https://conceptualreasoning.ai/dtbench via Epoch AI |
| Earthborne Rangers (EBR-bench) | Objectives reached | 28.6% | Independent | max | 7 Aug 2026 | Epoch AI |
| LMCA (Language Model Conceptual Argumentation) | Correlation with expert ratings (Pearson r) | 0.5751 | Third party | max | — | https://conceptualreasoning.ai/lmca via Epoch AI |
| Mystery Game Puzzles | Puzzles solved | 36.0% | Independent | max | 25 Jul 2026 | Epoch AI |
| Mystery Game Puzzles | Puzzles solved | 31.0% | Independent | xhigh | 26 Jul 2026 | Epoch AI |
| SimpleBench | Score (AVG@5) | 64.8% | Third party | unknown | — | https://lmcouncil.ai/benchmarks via Epoch AI |
Scientific reasoning
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| GPQA Diamond | Accuracy | 91.0% | Independent | max | 7 Jun 2026 | Epoch AI |
| GPQA Diamond | Accuracy | 88.4% | Independent | low | 6 Aug 2026 | Epoch AI |
| GPQA Diamond | Accuracy | 85.4% | Independent | none | 6 Aug 2026 | Epoch AI |
| Surface Evolver Bench | Mean score (partial credit) | 87.5% | Third party | high | — | https://yhenon.github.io/surface-evolver-llm-eval/ via Epoch AI |
| Surface Evolver Bench | Mean score (partial credit) | 68.1% | Third party | none | — | https://yhenon.github.io/surface-evolver-llm-eval/ via Epoch AI |
Mathematics
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| FrontierMath Tier 4 | Accuracy | 56.1% | Independent | max | 10 Jun 2026 | Epoch AI |
| FrontierMath Tier 4 (v1) | Accuracy | 31.3% | Independent | max | 8 Jun 2026 | Epoch AI |
| FrontierMath Tiers 1-3 | Accuracy | 80.0% | Independent | max | 10 Jun 2026 | Epoch AI |
| FrontierMath Tiers 1-3 (v1) | Accuracy | 47.2% | Independent | max | 8 Jun 2026 | Epoch AI |
| OTIS Mock AIME 2024-2025 | Accuracy | 98.3% | Independent | max | 7 Jun 2026 | Epoch AI |
| OTIS Mock AIME 2024-2025 | Accuracy | 97.8% | Independent | low | 6 Aug 2026 | Epoch AI |
| OTIS Mock AIME 2024-2025 | Accuracy | 84.4% | Independent | none | 6 Aug 2026 | Epoch AI |
Coding & software engineering
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| DeepSWE | Pass@1 | 59.0% | Third party | max | — | https://deepswe.datacurve.ai/ via Epoch AI |
| DeepSWE | Pass@1 | 54.4% | Third party | xhigh | — | https://deepswe.datacurve.ai/ via Epoch AI |
| DeepSWE | Pass@1 | 51.8% | Third party | high | — | https://deepswe.datacurve.ai/ via Epoch AI |
| DeepSWE | Pass@1 | 48.7% | Third party | medium | — | https://deepswe.datacurve.ai/ via Epoch AI |
| DeepSWE | Pass@1 | 40.8% | Third party | low | — | https://deepswe.datacurve.ai/ via Epoch AI |
| FrontierCode | Main score (Mean@5) | 46.5% | Third party | unknown | — | https://cognition.com/frontiercode via Epoch AI |
| GSO | OPT@1 | 47.1% | Third party | unknown | — | https://gso-bench.github.io/index.html via Epoch AI |
| PostTrainBench | Average score | 33.8% | Source not stated | high | — | Epoch AI via Epoch AI |
| PostTrainBench | Average score | 32.9% | Source not stated | max | — | Epoch AI via Epoch AI |
| WeirdML (v2) | Average accuracy | 82.9% | Third party | xhigh | — | https://htihle.github.io/weirdml.html via Epoch AI |
| WeirdML (v2) | Average accuracy | 76.0% | Third party | medium | — | https://htihle.github.io/weirdml.html via Epoch AI |
| WeirdML (v2) | Average accuracy | 70.5% | Third party | none | — | https://htihle.github.io/weirdml.html via Epoch AI |
Agentic tool use
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| Deep Research Bench (FutureSearch) | Average score | 50.2% | Third party | high | — | https://evals.futuresearch.ai/#drb via Epoch AI |
| Deep Research Bench (FutureSearch) | Average score | 49.3% | Third party | low | — | https://evals.futuresearch.ai/#drb via Epoch AI |
| Deep Research Bench (FutureSearch) | Average score | 47.4% | Third party | medium | — | https://evals.futuresearch.ai/#drb via Epoch AI |
Computer & browser use
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| OSWorld 2.0 | Binary completion (500 steps) | 20.6% | Third party | max | — | https://osworld-v2.xlang.ai/ via Epoch AI |
| OSWorld 2.0 | Binary completion (500 steps) | 18.5% | Third party | max | — | https://osworld-v2.xlang.ai/ via Epoch AI |
Professional knowledge work
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| APEX-Agents | Pass@1 (all rubric criteria met) | 48.9% | Source not stated | max | — | Epoch AI via Epoch AI |
Vision & multimodal understanding
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| Furniture Assembly | Score | 42.5% | Independent | max | 10 Sep 2026 | Epoch AI |
Factuality & hallucination
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| SimpleQA Verified | Accuracy (% correct) | 53.0% | Independent | max | 27 Aug 2026 | Epoch AI |
Human preference
| Benchmark | Metric | Value | Provenance | Setting | Evaluated | Source |
|---|---|---|---|---|---|---|
| Code Arena: WebDev (arena.ai) | Arena score (Elo scale) | 1555 | Independent | high | 25 Sep 2026 | Arena |
| Code Arena: WebDev (arena.ai) | Arena score (Elo scale) | 1535 | Independent | — | 25 Sep 2026 | Arena |
| Code Arena: WebDev (arena.ai) | Votes | 17611 | Independent | high | 25 Sep 2026 | Arena |
| Code Arena: WebDev (arena.ai) | Votes | 16576 | Independent | — | 25 Sep 2026 | Arena |
| Document Arena (arena.ai) | Arena score (Elo scale) | 1486 | Independent | high | 13 Sep 2026 | Arena |
| Document Arena (arena.ai) | Arena score (Elo scale) | 1475 | Independent | — | 13 Sep 2026 | Arena |
| Document Arena (arena.ai) | Votes | 12128 | Independent | — | 13 Sep 2026 | Arena |
| Document Arena (arena.ai) | Votes | 11551 | Independent | high | 13 Sep 2026 | Arena |
| Search Arena (arena.ai) | Arena score (Elo scale) | 1195 | Independent | — | 24 Aug 2026 | Arena |
| Search Arena (arena.ai) | Votes | 70998 | Independent | — | 24 Aug 2026 | Arena |
| Text Arena (arena.ai) | Arena score (Elo scale) | 1480 | Independent | high | 25 Sep 2026 | Arena |
| Text Arena (arena.ai) | Arena score (Elo scale) | 1474 | Independent | — | 25 Sep 2026 | Arena |
| Text Arena (arena.ai) | Votes | 62717 | Independent | — | 25 Sep 2026 | Arena |
| Text Arena (arena.ai) | Votes | 61560 | Independent | high | 25 Sep 2026 | Arena |
| Vision Arena (arena.ai) | Arena score (Elo scale) | 1286 | Independent | high | 28 Sep 2026 | Arena |
| Vision Arena (arena.ai) | Arena score (Elo scale) | 1279 | Independent | — | 28 Sep 2026 | Arena |
| Vision Arena (arena.ai) | Votes | 18583 | Independent | — | 28 Sep 2026 | Arena |
| Vision Arena (arena.ai) | Votes | 18020 | Independent | high | 28 Sep 2026 | Arena |
Other Claude Opus versions
- Claude Opus 5.522 Sep 2026
- Claude Opus 524 Jul 2026
- Claude Opus 4.716 Apr 2026
- Claude Opus 4.65 Feb 2026
- Claude Opus 4.524 Nov 2025
- Claude Opus 4.15 Aug 2025
- Claude Opus 422 May 2025
- Claude 3 Opus4 Mar 2024