Rankings / Anthropic
Claude Opus 4
release date unknown
BenchAtlas Index
as of 2026-07-14
thinking
rank #245
4 families · 4 categories · medium
base configuration
rank #294
4 families · 4 categories · medium
Benchmark evidence
10 results
coding
| LiveCodeBench v6 problems=454 · window_start=2024-08-01 | thinking | 56.6 % | independent | LiveCodeBench Leaderboard |
| LiveCodeBench v6 problems=454 · window_start=2024-08-01 | 46.9 % | independent | LiveCodeBench Leaderboard |
factuality
| MASK | thinking | 87.9 % 84.1–91.6 | independent | Scale Labs |
| MASK contamination=Potential contamination warning: This model was evaluated after the public release of MASK, allowing model builder access to the prompts and solutions. | 80.3 % 79.7–80.9 | independent | Scale Labs |
knowledge science
| Humanity's Last Exam contamination=Potential contamination warning: This model was evaluated after the public release of HLE, allowing model builder access to the prompts and solutions. · implementation=scale | thinking | 10.7 % 9.5–11.9 | independent | Scale Labs |
| Humanity's Last Exam contamination=Potential contamination warning: This model was evaluated after the public release of HLE, allowing model builder access to the prompts and solutions. · implementation=scale | 6.7 % 5.7–7.7 | independent | Scale Labs |
multimodal
| VISTA | thinking | 47.0 % 46.0–47.9 | independent | Scale Labs |
| VISTA | 43.5 % 42.3–44.8 | independent | Scale Labs |
reasoning math
| EnigmaEval contamination=Potential contamination warning: This model was evaluated after the public release of EnigmaEval, allowing model builder access to the prompts and solutions. | thinking | 5.6 % 4.3–6.9 | independent | Scale Labs |
| EnigmaEval contamination=Potential contamination warning: This model was evaluated after the public release of EnigmaEval, allowing model builder access to the prompts and solutions. | 3.2 % 2.2–4.2 | independent | Scale Labs |
Agent + model results
systems, not bare-model scores
| agent + model goose + Claude Opus 4 | Terminal-Bench 1.0 | 45.3 % | unverified | Terminal-Bench Leaderboard |
| agent + model Claude Code + Claude Opus 4 | Terminal-Bench 1.0 | 43.2 % | unverified | Terminal-Bench Leaderboard |
| agent + model goose + Claude Opus 4 | Terminal-Bench 1.0 | 42.0 % | unverified | Terminal-Bench Leaderboard |
| agent + model terminus-2 + Claude Opus 4 | Terminal-Bench 1.0 | 39.0 % | community | Terminal-Bench Leaderboard |
These scores measure the whole agent system (scaffold, tools, budgets) — they are never merged into the bare model’s numbers.