Rankings / Anthropic
Claude Opus 4.5 20251101
released 2025-11-24
BenchAtlas Index
as of 2026-07-14
high effort
rank #79
7 families · 4 categories · medium
thinking
rank #96
6 families · 5 categories · medium
high effort
rank #107
7 families · 4 categories · medium
Benchmark evidence
115 results
agentic coding
| Release 2026-06-25 tasks_counted=3 · livebench_version=2026-06-25 | high effort · 64K budget | 39.7 % | independent | LiveBench |
coding
| Release 2026-06-25 tasks_counted=2 · livebench_version=2026-06-25 | high effort · 64K budget | 79.7 % | independent | LiveBench |
data analysis
| Release 2026-06-25 tasks_counted=3 · livebench_version=2026-06-25 | high effort · 64K budget | 74.4 % | independent | LiveBench |
external indices
| ECI | 149.5 points | independent | Epoch AI Benchmarking Hub | |
| ECI | 32K budget | 149.5 points | independent | Epoch AI Benchmarking Hub |
| ECI | 8K budget | 149.5 points | independent | Epoch AI Benchmarking Hub |
| ECI | 64K budget | 149.5 points | independent | Epoch AI Benchmarking Hub |
| ECI | 128K budget | 149.5 points | independent | Epoch AI Benchmarking Hub |
| ECI | 16K budget | 149.5 points | independent | Epoch AI Benchmarking Hub |
factuality
| MASK contamination=Potential contamination warning: This model was evaluated after the public release of MASK, allowing model builder access to the prompts and solutions. | thinking | 92.5 % 91.3–93.8 | 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. | 87.1 % 86.1–88.2 | independent | Scale Labs | |
| SimpleQA Verified | 32K budget | 41.8 % | independent | Epoch AI Benchmarking Hub 2025-12-09 |
human preference
| Coding (style control)older version arena=text · category=coding · style_control=true | 1522.6 1517.1–1528.2 | community | LMArena Leaderboard Dataset | |
| Industry Software And It Services (style control)older version arena=text · category=industry_software_and_it_services · style_control=true | 1506.7 1502.0–1511.5 | community | LMArena Leaderboard Dataset | |
| Expert (style control)older version arena=text · category=expert · style_control=true | 1502.8 1493.9–1511.8 | community | LMArena Leaderboard Dataset | |
| Chinese (style control)older version arena=text · category=chinese · style_control=true | 1500.4 1489.7–1511.2 | community | LMArena Leaderboard Dataset | |
| French (style control)older version arena=text · category=french · style_control=true | 1499.2 1483.1–1515.3 | community | LMArena Leaderboard Dataset | |
| Hard Prompts English (style control)older version arena=text · category=hard_prompts_english · style_control=true | 1498.5 1493.3–1503.7 | community | LMArena Leaderboard Dataset | |
| Hard Prompts (style control)older version arena=text · category=hard_prompts · style_control=true | 1497.8 1493.7–1501.9 | community | LMArena Leaderboard Dataset | |
| Longer Query (style control)older version arena=text · category=longer_query · style_control=true | 1492.4 1487.3–1497.6 | community | LMArena Leaderboard Dataset | |
| Industry Life And Physical And Social Science (style control)older version arena=text · category=industry_life_and_physical_and_social_science · style_control=true | 1487.7 1481.3–1494.0 | community | LMArena Leaderboard Dataset | |
| Industry Medicine And Healthcare (style control)older version arena=text · category=industry_medicine_and_healthcare · style_control=true | 1486.7 1477.3–1496.2 | community | LMArena Leaderboard Dataset | |
| Industry Legal And Government (style control)older version arena=text · category=industry_legal_and_government · style_control=true | 1484.0 1475.1–1492.8 | community | LMArena Leaderboard Dataset | |
| Multi Turn (style control)older version arena=text · category=multi_turn · style_control=true | 1483.9 1477.7–1490.1 | community | LMArena Leaderboard Dataset | |
| Industry Mathematical (style control)older version arena=text · category=industry_mathematical · style_control=true | 1479.1 1468.4–1489.8 | community | LMArena Leaderboard Dataset | |
| English (style control)older version arena=text · category=english · style_control=true | 1478.6 1474.3–1482.9 | community | LMArena Leaderboard Dataset | |
| Exclude Ties (style control)older version arena=text · category=exclude_ties · style_control=true | 1477.8 1473.5–1482.2 | community | LMArena Leaderboard Dataset | |
| Instruction Following (style control)older version arena=text · category=instruction_following · style_control=true | 1475.1 1469.9–1480.2 | community | LMArena Leaderboard Dataset | |
| Russian (style control)older version arena=text · category=russian · style_control=true | 1473.5 1466.1–1481.0 | community | LMArena Leaderboard Dataset | |
| Industry Business And Management And Financial Operations (style control)older version arena=text · category=industry_business_and_management_and_financial_operations · style_control=true | 1472.0 1465.9–1478.0 | community | LMArena Leaderboard Dataset | |
| German (style control)older version arena=text · category=german · style_control=true | 1471.7 1452.6–1490.8 | community | LMArena Leaderboard Dataset | |
| Overall (style control) arena=text · category=overall · style_control=true | 1469.2 1465.9–1472.4 | community | LMArena Leaderboard Dataset | |
| Spanish (style control)older version arena=text · category=spanish · style_control=true | 1466.9 1451.8–1482.0 | community | LMArena Leaderboard Dataset | |
| Polish (style control)older version arena=text · category=polish · style_control=true | 1466.7 1452.3–1481.0 | community | LMArena Leaderboard Dataset | |
| Math (style control)older version arena=text · category=math · style_control=true | 1465.4 1456.1–1474.7 | community | LMArena Leaderboard Dataset | |
| Industry Writing And Literature And Language (style control)older version arena=text · category=industry_writing_and_literature_and_language · style_control=true | 1464.8 1459.2–1470.4 | community | LMArena Leaderboard Dataset | |
| Creative Writing (style control)older version arena=text · category=creative_writing · style_control=true | 1460.8 1454.3–1467.4 | community | LMArena Leaderboard Dataset | |
| Non English (style control)older version arena=text · category=non_english · style_control=true | 1455.9 1451.8–1460.0 | community | LMArena Leaderboard Dataset | |
| Industry Entertainment And Sports And Media (style control)older version arena=text · category=industry_entertainment_and_sports_and_media · style_control=true | 1452.9 1446.9–1458.9 | community | LMArena Leaderboard Dataset | |
| Korean (style control)older version arena=text · category=korean · style_control=true | 1449.1 1429.9–1468.4 | community | LMArena Leaderboard Dataset | |
| Japanese (style control)older version arena=text · category=japanese · style_control=true | 1443.5 1417.3–1469.6 | community | LMArena Leaderboard Dataset |
knowledge science
| GPQA Diamond | 32K budget | 86.0 % | independent | Epoch AI Benchmarking Hub 2025-11-24 |
| GPQA Diamond implementation=vals-ai | high effort | 85.9 % | independent | Vals AI |
| GPQA Diamond | 16K budget | 85.5 % | independent | Epoch AI Benchmarking Hub 2025-11-25 |
| GPQA Diamond | 80.7 % | independent | Epoch AI Benchmarking Hub 2025-11-24 | |
| GPQA Diamond implementation=vals-ai | high effort | 79.5 % | independent | Vals AI |
| 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 | 25.2 % 23.5–26.9 | 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 | 14.2 % 12.8–15.5 | independent | Scale Labs | |
| MMLU-Pro implementation=vals-ai | high effort | 87.3 % | independent | Vals AI |
| MMLU-Pro implementation=vals-ai | high effort | 85.6 % | independent | Vals AI |
language
| Release 2026-06-25 tasks_counted=3 · livebench_version=2026-06-25 | high effort · 64K budget | 81.3 % | independent | LiveBench |
long context instruction
| Release 2026-06-25 tasks_counted=4 · livebench_version=2026-06-25 | high effort · 64K budget | 62.5 % | independent | LiveBench |
| MultiChallenge | thinking | 59.0 % 58.5–59.4 | independent | Scale Labs |
multimodal
| MMMU implementation=vals-ai | high effort | 82.9 % | independent | Vals AI |
| MMMU implementation=vals-ai | high effort | 81.1 % | independent | Vals AI |
| VISTA | thinking | 46.4 % 46.4–46.5 | independent | Scale Labs |
| VISTA | 45.3 % 45.3–45.4 | independent | Scale Labs |
professional
| CaseLaw | thinking | 62.6 % | independent | Vals AI |
| CorpFin | high effort | 65.1 % | independent | Vals AI |
| CorpFin | high effort | 61.3 % | independent | Vals AI |
| LegalBench | high effort | 84.6 % | independent | Vals AI |
| LegalBench | high effort | 82.8 % | independent | Vals AI |
| MedQA | thinking | 95.9 % | independent | Vals AI |
| MedQA | 93.2 % | independent | Vals AI | |
| TaxEval | high effort | 74.9 % | independent | Vals AI |
| TaxEval | high effort | 74.3 % | independent | Vals AI |
reasoning math
| AIME implementation=vals-ai | thinking | 95.4 % | independent | Vals AI |
| AIME implementation=vals-ai | 76.9 % | independent | Vals AI | |
| ARC-AGI-1older version split=public_eval · model_type=CoT | thinking · 32K budget | 86.6 % | independent | ARC Prize Leaderboard |
| ARC-AGI-1older version split=public_eval · model_type=CoT | thinking · 16K budget | 81.6 % | independent | ARC Prize Leaderboard |
| ARC-AGI-1older version split=semi_private · model_type=CoT | thinking · 64K budget | 80.0 % | independent | ARC Prize Leaderboard |
| ARC-AGI-1older version split=semi_private · model_type=CoT | thinking · 32K budget | 75.8 % | independent | ARC Prize Leaderboard |
| ARC-AGI-1older version split=semi_private · model_type=CoT | thinking · 16K budget | 72.0 % | independent | ARC Prize Leaderboard |
| ARC-AGI-1older version split=public_eval · model_type=CoT | thinking · 8K budget | 70.1 % | independent | ARC Prize Leaderboard |
| ARC-AGI-1older version split=semi_private · model_type=CoT | thinking · 8K budget | 58.7 % | independent | ARC Prize Leaderboard |
| ARC-AGI-1older version split=public_eval · model_type=CoT | no reasoning | 52.6 % | independent | ARC Prize Leaderboard |
| ARC-AGI-1older version split=semi_private · model_type=CoT | no reasoning | 40.0 % | independent | ARC Prize Leaderboard |
| ARC-AGI-2 split=semi_private · model_type=CoT | thinking · 64K budget | 37.6 % | independent | ARC Prize Leaderboard |
| ARC-AGI-2 split=public_eval · model_type=CoT | thinking · 32K budget | 28.1 % | independent | ARC Prize Leaderboard |
| ARC-AGI-2 split=public_eval · model_type=CoT | thinking · 16K budget | 24.2 % | independent | ARC Prize Leaderboard |
| ARC-AGI-2 split=semi_private · model_type=CoT | thinking · 16K budget | 22.8 % | independent | ARC Prize Leaderboard |
| ARC-AGI-2 split=semi_private · model_type=CoT | thinking · 8K budget | 13.9 % | independent | ARC Prize Leaderboard |
| ARC-AGI-2 split=public_eval · model_type=CoT | thinking · 8K budget | 10.4 % | independent | ARC Prize Leaderboard |
| ARC-AGI-2 split=semi_private · model_type=CoT | no reasoning | 7.8 % | independent | ARC Prize Leaderboard |
| ARC-AGI-2 split=public_eval · model_type=CoT | no reasoning | 7.1 % | independent | ARC Prize Leaderboard |
| ARC-AGI-2 split=semi_private · model_type=CoT | thinking · 64K budget | 2.4 usd_per_task | independent | ARC Prize Leaderboard |
| ARC-AGI-1older version split=semi_private · model_type=CoT | thinking · 64K budget | 1.5 usd_per_task | independent | ARC Prize Leaderboard |
| ARC-AGI-2 split=public_eval · model_type=CoT | thinking · 32K budget | 1.4 usd_per_task | independent | ARC Prize Leaderboard |
| ARC-AGI-1older version split=semi_private · model_type=CoT | thinking · 32K budget | 0.9 usd_per_task | independent | ARC Prize Leaderboard |
| ARC-AGI-2 split=public_eval · model_type=CoT | thinking · 16K budget | 0.8 usd_per_task | independent | ARC Prize Leaderboard |
| ARC-AGI-2 split=semi_private · model_type=CoT | thinking · 16K budget | 0.8 usd_per_task | independent | ARC Prize Leaderboard |
| ARC-AGI-1older version split=public_eval · model_type=CoT | thinking · 32K budget | 0.8 usd_per_task | independent | ARC Prize Leaderboard |
| ARC-AGI-1older version split=semi_private · model_type=CoT | thinking · 16K budget | 0.6 usd_per_task | independent | ARC Prize Leaderboard |
| ARC-AGI-1older version split=public_eval · model_type=CoT | thinking · 16K budget | 0.5 usd_per_task | independent | ARC Prize Leaderboard |
| ARC-AGI-2 split=public_eval · model_type=CoT | thinking · 8K budget | 0.5 usd_per_task | independent | ARC Prize Leaderboard |
| ARC-AGI-2 split=semi_private · model_type=CoT | thinking · 8K budget | 0.5 usd_per_task | independent | ARC Prize Leaderboard |
| ARC-AGI-1older version split=semi_private · model_type=CoT | thinking · 8K budget | 0.4 usd_per_task | independent | ARC Prize Leaderboard |
| ARC-AGI-1older version split=public_eval · model_type=CoT | thinking · 8K budget | 0.3 usd_per_task | independent | ARC Prize Leaderboard |
| ARC-AGI-2 split=public_eval · model_type=CoT | no reasoning | 0.2 usd_per_task | independent | ARC Prize Leaderboard |
| ARC-AGI-2 split=semi_private · model_type=CoT | no reasoning | 0.2 usd_per_task | independent | ARC Prize Leaderboard |
| ARC-AGI-1older version split=semi_private · model_type=CoT | no reasoning | 0.1 usd_per_task | independent | ARC Prize Leaderboard |
| ARC-AGI-1older version split=public_eval · model_type=CoT | no reasoning | 0.1 usd_per_task | independent | ARC Prize Leaderboard |
| 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 | 11.9 % 10.1–13.8 | 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. | 4.7 % 3.5–5.8 | independent | Scale Labs | |
| FrontierMath | 20.7 % | independent | Epoch AI Benchmarking Hub 2025-11-25 | |
| FrontierMath | 32K budget | 20.7 % | independent | Epoch AI Benchmarking Hub 2025-11-25 |
| FrontierMath | 16K budget | 20.3 % | independent | Epoch AI Benchmarking Hub 2025-11-25 |
| FrontierMath Tier 4older version | 32K budget | 4.2 % | independent | Epoch AI Benchmarking Hub 2025-11-25 |
| FrontierMath Tier 4older version | 4.2 % | independent | Epoch AI Benchmarking Hub 2025-11-25 | |
| FrontierMath Tier 4older version | 16K budget | 2.1 % | independent | Epoch AI Benchmarking Hub 2025-11-25 |
| Release 2026-06-25 tasks_counted=4 · livebench_version=2026-06-25 | high effort · 64K budget | 90.4 % | independent | LiveBench |
| Release 2026-06-25 tasks_counted=4 · livebench_version=2026-06-25 | high effort · 64K budget | 80.1 % | independent | LiveBench |
| OTIS Mock AIME 2024–2025 | 32K budget | 86.1 % | independent | Epoch AI Benchmarking Hub 2025-11-24 |
| OTIS Mock AIME 2024–2025 | 16K budget | 81.7 % | independent | Epoch AI Benchmarking Hub 2025-11-24 |
| OTIS Mock AIME 2024–2025 | 48.1 % | independent | Epoch AI Benchmarking Hub 2025-11-24 |
Agent + model results
systems, not bare-model scores
| agent + model live-SWE-agent + Claude Opus 4.5 20251101 | SWE-bench Verified | 79.2 % | unverified | SWE-bench Leaderboard |
| agent + model Epoch Inspect harness + Claude Opus 4.5 20251101 | SWE-bench Verified | 76.7 % | independent | Epoch AI Benchmarking Hub |
| agent + model mini-SWE-agent + Claude Opus 4.5 20251101 | SWE-bench bash-only | 74.4 % | community | SWE-bench Leaderboard |
| agent + model mini-SWE-agent + Claude Opus 4.5 20251101 | SWE-bench Verified | 74.4 % | community | SWE-bench Leaderboard |
| agent + model mini-SWE-agent + Claude Opus 4.5 20251101 | SWE-bench Multilingual | 70.7 % | community | SWE-bench Leaderboard |
| agent + model mini-SWE-agent + Claude Opus 4.5 20251101 | SWE-bench Multilingual | 0.8 usd_per_task | community | SWE-bench Leaderboard |
| agent + model mini-SWE-agent + Claude Opus 4.5 20251101 | SWE-bench Verified | 0.7 usd_per_task | community | SWE-bench Leaderboard |
| agent + model mini-SWE-agent + Claude Opus 4.5 20251101 | SWE-bench bash-only | 0.7 usd_per_task | community | SWE-bench Leaderboard |
| agent + model Droid + Claude Opus 4.5 20251101 | Terminal-Bench 2.0 | 63.1 % | unverified | Terminal-Bench Leaderboard |
| agent + model terminus-2 + Claude Opus 4.5 20251101 | Terminal-Bench 2.0 | 57.8 % | community | Terminal-Bench Leaderboard |
| agent + model Claude Code + Claude Opus 4.5 20251101 | Terminal-Bench 2.0 | 52.1 % | community | Terminal-Bench Leaderboard |
| agent + model OpenHands + Claude Opus 4.5 20251101 | Terminal-Bench 2.0 | 51.9 % | community | Terminal-Bench Leaderboard |
These scores measure the whole agent system (scaffold, tools, budgets) — they are never merged into the bare model’s numbers.