Anthropic
Claude Opus 4.7
Opus upgrade with the top score in Epoch AI's SWE-bench Verified runs (83.5%).
- Announced
- Access
- Public
Benchmark scores
| Test | Score | Best on the tracker | How it was measured |
|---|---|---|---|
| Epoch Capabilities Indexmany tests combined into one score | 158 points | GPT-6.2 Astra 174 points | Independentlisted as "Claude Opus 4.7", CI 155.69 to 159.63Run 9 Nov 2026 |
| MirrorCoderebuilding whole programs | 31% | Claude Opus 5.5 79% | Independenthigh reasoning effort, MirrorCode (ML, +Private, 2L): 15 programs x 2 languages, 3 runs each, 10B-token / 7-day budget, ±9.0 SERun 12 Aug 2026 |
| APEX-Agentsprofessional office tasks | 49% | Gemini 4 Carbon 84% | Independentpass@1, max effort; listed as "Opus 4.7" |
| Humanity's Last Examexpert questions, no tools | 36% | GPT-6 Astra 55% | Independentleaderboard entry "claude-opus-4-7" (claude-opus-4-7_unknown) |
| ARC-AGI-3puzzle games, standard harnessTypical first-time human player: 100% | 0.2% | Claude Fable 5.2 81% | IndependentARC Prize config "Opus 4.7 (High)" (anthropic-opus-4-7-high), standard harness |
| FrontierMath T1–3hard maths problems | 70% | Claude Fable 5.2 95% | Independentmax reasoning effort (claude-opus-4-7_max)Run 10 Jun 2026 |
| GPQA DiamondPhD-level science questionsPhD experts: 70%Saturated | 90% | Gemini 4 Argon 96% | Independentxhigh reasoning effort (claude-opus-4-7_xhigh)Run 17 Apr 2026 |
| FrontierMath T4research-level maths problemsSaturated | 32% | GPT-6.1 Sol 100% | Independentmax reasoning effort (claude-opus-4-7_max)Run 10 Jun 2026 |
| ARC-AGI-2visual puzzles, model aloneAverage person: 60%Saturated | 76% | GPT-6 Astra 95% | IndependentARC Prize config "Claude 4.7 (Max)" (claude-4-7-max) |
| OTIS Mock AIMEcompetition mathsSaturated | 98% | GPT-5.5 100% | Independentxhigh reasoning effort (claude-opus-4-7_xhigh)Run 17 Apr 2026 |
| SWE-bench Verifiedfixing real software bugsKnown flaws | 83% | Top score | Independentmax reasoning effort (claude-opus-4-7_max)Run 20 Apr 2026 |
| METR time horizon (TH1.1 + TH2)task length an AI agent finishes half the time | 8.4 h | GPT-6 Astra 1.7 days | IndependentPooled TH1.1 + TH2 suites; 95% CI 257.9–937.6 min |
Also on these charts
Sources
- Introducing Claude Opus 4.7Anthropic · Official · anthropic.com
- SWE-bench Verified (Epoch)Epoch AI · Dataset · epoch.ai
- Epoch AI Benchmarking Hub (model release dates and runs)Epoch AI · Dataset · 9 Nov 2026 · epoch.ai