Claude Opus 4.6 vs Claude Opus 4.7
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models
Opus 4.619.1
Opus 4.733.3
Arena Text · overall Arena Text · overall: Human pairwise preferences in the stated text vote pool
Opus 4.61,503
Opus 4.71,501
Arena Text · Chinese Arena Text · chinese: Human pairwise preferences in the stated text vote pool
Opus 4.61,550.2
Opus 4.71,531.8
Arena Text · English Arena Text · english: Human pairwise preferences in the stated text vote pool
Opus 4.61,512
Opus 4.71,496.3
Arena Text · French Arena Text · french: Human pairwise preferences in the stated text vote pool
Opus 4.61,512.8
Opus 4.71,503
Arena Text · German Arena Text · german: Human pairwise preferences in the stated text vote pool
Opus 4.61,501.8
Opus 4.71,495.2
Arena Text · instruction following Arena Text · instruction following: Human pairwise preferences in the stated text vote pool
Opus 4.61,522
Opus 4.71,498
Arena Text · Japanese Arena Text · japanese: Human pairwise preferences in the stated text vote pool
Opus 4.61,488
Opus 4.71,473.9
Arena Text · Korean Arena Text · korean: Human pairwise preferences in the stated text vote pool
Opus 4.61,449.8
Opus 4.71,450.9
Arena Text · non-English Arena Text · non english: Human pairwise preferences in the stated text vote pool
Opus 4.61,488.1
Opus 4.71,480.5
Arena Text · Polish Arena Text · polish: Human pairwise preferences in the stated text vote pool
Opus 4.61,496.6
Opus 4.71,479.5
Arena Text · Russian Arena Text · russian: Human pairwise preferences in the stated text vote pool
Opus 4.61,496.1
Opus 4.71,494.4
Arena Text · Spanish Arena Text · spanish: Human pairwise preferences in the stated text vote pool
Opus 4.61,504.2
Opus 4.71,494.6
AA Omniscience Index AA Omniscience Index: Scores factual knowledge while penalizing incorrect answers to encourage admitting uncertainty
Opus 4.62.4
Opus 4.714.8
SimpleQA Verified v1 SimpleQA Verified v1: Accuracy on short factual questions with verified answers, answered without web access
Opus 4.6 · max47
Opus 4.7 · xhigh51.7
CritPt (physics) CritPt (physics): Research-level physics problems that need many careful reasoning steps
Opus 4.62.8
Opus 4.75.1
AA Intelligence Index 4.3.2 AA Intelligence Index 4.3.2: Artificial Analysis combines tests of knowledge, reasoning, coding and tool use into one score
Opus 4.626.4
Opus 4.730.9
AA Omniscience · Accuracy AA Omniscience · Accuracy: Measures correct answers on the AA Omniscience factual knowledge test
Opus 4.645.8
Opus 4.744.7
AA Omniscience · Hallucination Rate AA Omniscience · Hallucination Rate: Measures how often attempted answers are incorrect on the AA Omniscience factual knowledge test · lower is better
Opus 4.680.1
Opus 4.754.1
AA-LCR (long context) 1.1 AA-LCR (long context) 1.1: Questions that need reading and connecting facts across very long documents (around 100K tokens)
Opus 4.667
Opus 4.775.7
IFBench (instructions) IFBench (instructions): Checks whether the model follows precise formatting and content instructions
Opus 4.644.6
Opus 4.743.6
EQ-Bench 3 EQ-Bench 3: Judge-rated responses to social and emotionally nuanced scenarios, scored through the source pairwise ranking
Opus 4.61,717.4
Opus 4.71,884.3
Chess Puzzles Chess Puzzles: Models choose the best next move from a novel chess position
Opus 4.6 · max21
Opus 4.7 · max41
FrontierMath FrontierMath: Models solve expert-written advanced mathematics problems on the named tier and edition
Opus 4.6 · max40.7
Opus 4.7 · xhigh43.8
FrontierMath Tier 4 FrontierMath Tier 4: Models solve expert-written advanced mathematics problems on the named tier and edition
Opus 4.6 · max22.9
Opus 4.7 · xhigh22.9
FrontierMath Tier 4 (v2) FrontierMath Tier 4 (v2): Models solve expert-written advanced mathematics problems on the named tier and edition
Opus 4.6 · max26.8
Opus 4.7 · max31.7
FrontierMath Tiers 1–3 (v2) FrontierMath Tiers 1–3 (v2): Models solve expert-written advanced mathematics problems on the named tier and edition
Opus 4.6 · max66
Opus 4.7 · max70.2
Mystery Game Puzzles Mystery Game Puzzles: Models choose the best move in puzzle positions from an undisclosed game
Opus 4.6 · max25
Opus 4.7 · max28
MathArena · Apex MathArena · apex · Apex: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.634.5
Opus 4.7 · xhigh40.6
MathArena · Apex Shortlist MathArena · apex · Apex Shortlist: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.686.7
Opus 4.7 · xhigh63.8
MathArena · ArXivMath · February 2026 MathArena · arxiv · 02/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.640.6
Opus 4.7 · xhigh40.6
MathArena · ArXivMath · January 2026 MathArena · arxiv · 01/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.672.8
Opus 4.7 · xhigh52.2
MathArena · ArXivMath · March 2026 MathArena · arxiv · 03/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.662.5
Opus 4.7 · xhigh50.8
MathArena · BrokenArXiv · February 2026 MathArena · arxiv_false · 02/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.63.2
Opus 4.7 · xhigh4
MathArena · BrokenArXiv · March 2026 MathArena · arxiv_false · 03/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.65.8
Opus 4.7 · xhigh5.8
Vals LegalBench Vals LegalBench: Legal reasoning and legal-language understanding on LegalBench
Opus 4.685.3
Opus 4.785.3
Vals MedCode Vals MedCode: Medical coding accuracy on patient encounters
Opus 4.649.1
Opus 4.754.9
Vals MedScribe Vals MedScribe: Accuracy of clinical notes produced from consultation transcripts
Opus 4.686.1
Opus 4.783
MLCR-AA MLCR-AA: Models synthesize long medical records; the headline credits concise answers judged both accurate and complete
Opus 4.6—
Opus 4.728.9
ARC-AGI-1 · public evaluation ARC-AGI-1 · public-eval: Models infer transformations from examples and solve unseen grid puzzles; the evaluation split is explicit
Opus 4.696.3
Opus 4.7—
ARC-AGI-1 · semi-private ARC-AGI-1 · semi-private: Models infer transformations from examples and solve unseen grid puzzles; the evaluation split is explicit
Opus 4.694
Opus 4.7—
ARC-AGI-2 · public evaluation ARC-AGI-2 · public-eval: Models infer transformations from examples and solve unseen grid puzzles; the evaluation split is explicit
Opus 4.679
Opus 4.7—
ARC-AGI-2 · semi-private ARC-AGI-2 · semi-private: Models infer transformations from examples and solve unseen grid puzzles; the evaluation split is explicit
Opus 4.669.2
Opus 4.7—
MathArena · ArXivMath · April 2026 MathArena · arxiv · 04/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.6—
Opus 4.7 · xhigh57.5
MathArena · ArXivMath · December 2025 MathArena · arxiv · 12/2025: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.657.4
Opus 4.7—
MathArena · BrokenArXiv · April 2026 MathArena · arxiv_false · 04/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.6—
Opus 4.7 · xhigh4.1
MathArena · ArXivLean · March 2026 MathArena · arxivlean · 03/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.6—
Opus 4.714.6
MathArena · Project Euler MathArena · euler · Project Euler: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.687.7
Opus 4.7—
MathArena · Kangaroo 2025 1-2 MathArena · kangaroo · Kangaroo 2025 1-2: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.659.4
Opus 4.7—
MathArena · Kangaroo 2025 11-12 MathArena · kangaroo · Kangaroo 2025 11-12: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.687.5
Opus 4.7—
MathArena · Kangaroo 2025 3-4 MathArena · kangaroo · Kangaroo 2025 3-4: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.650
Opus 4.7—
MathArena · Kangaroo 2025 5-6 MathArena · kangaroo · Kangaroo 2025 5-6: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.658.3
Opus 4.7—
MathArena · Kangaroo 2025 7-8 MathArena · kangaroo · Kangaroo 2025 7-8: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.686.7
Opus 4.7—
MathArena · Kangaroo 2025 9-10 MathArena · kangaroo · Kangaroo 2025 9-10: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.691.7
Opus 4.7—
MathArena · USAMO 2026 MathArena · usamo · USAMO 2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.647
Opus 4.7—
Vals Legal Research Bench Vals Legal Research Bench: Grounded legal research; a task passes only if all criteria pass
Opus 4.6—
Opus 4.738.5
| Benchmark | Opus 4.6 | Opus 4.7 |
|---|---|---|
| Humanity's Last Exam | 19.1 | 33.3 |
| Arena Text · overall | 1,503 | 1,501 |
| Arena Text · Chinese | 1,550.2 | 1,531.8 |
| Arena Text · English | 1,512 | 1,496.3 |
| Arena Text · French | 1,512.8 | 1,503 |
| Arena Text · German | 1,501.8 | 1,495.2 |
| Arena Text · instruction following | 1,522 | 1,498 |
| Arena Text · Japanese | 1,488 | 1,473.9 |
| Arena Text · Korean | 1,449.8 | 1,450.9 |
| Arena Text · non-English | 1,488.1 | 1,480.5 |
| Arena Text · Polish | 1,496.6 | 1,479.5 |
| Arena Text · Russian | 1,496.1 | 1,494.4 |
| Arena Text · Spanish | 1,504.2 | 1,494.6 |
| AA Omniscience Index | 2.4 | 14.8 |
| SimpleQA Verified v1 | 47 · max | 51.7 · xhigh |
| CritPt (physics) | 2.8 | 5.1 |
| AA Intelligence Index 4.3.2 | 26.4 | 30.9 |
| AA Omniscience · Accuracy | 45.8 | 44.7 |
| AA Omniscience · Hallucination Rate | 80.1 | 54.1 |
| AA-LCR (long context) 1.1 | 67 | 75.7 |
| IFBench (instructions) | 44.6 | 43.6 |
| EQ-Bench 3 | 1,717.4 | 1,884.3 |
| Chess Puzzles | 21 · max | 41 · max |
| FrontierMath | 40.7 · max | 43.8 · xhigh |
| FrontierMath Tier 4 | 22.9 · max | 22.9 · xhigh |
| FrontierMath Tier 4 (v2) | 26.8 · max | 31.7 · max |
| FrontierMath Tiers 1–3 (v2) | 66 · max | 70.2 · max |
| Mystery Game Puzzles | 25 · max | 28 · max |
| MathArena · Apex | 34.5 | 40.6 · xhigh |
| MathArena · Apex Shortlist | 86.7 | 63.8 · xhigh |
| MathArena · ArXivMath · February 2026 | 40.6 | 40.6 · xhigh |
| MathArena · ArXivMath · January 2026 | 72.8 | 52.2 · xhigh |
| MathArena · ArXivMath · March 2026 | 62.5 | 50.8 · xhigh |
| MathArena · BrokenArXiv · February 2026 | 3.2 | 4 · xhigh |
| MathArena · BrokenArXiv · March 2026 | 5.8 | 5.8 · xhigh |
| Vals LegalBench | 85.3 | 85.3 |
| Vals MedCode | 49.1 | 54.9 |
| Vals MedScribe | 86.1 | 83 |
| MLCR-AA | — | 28.9 |
| ARC-AGI-1 · public evaluation | 96.3 | — |
| ARC-AGI-1 · semi-private | 94 | — |
| ARC-AGI-2 · public evaluation | 79 | — |
| ARC-AGI-2 · semi-private | 69.2 | — |
| MathArena · ArXivMath · April 2026 | — | 57.5 · xhigh |
| MathArena · ArXivMath · December 2025 | 57.4 | — |
| MathArena · BrokenArXiv · April 2026 | — | 4.1 · xhigh |
| MathArena · ArXivLean · March 2026 | — | 14.6 |
| MathArena · Project Euler | 87.7 | — |
| MathArena · Kangaroo 2025 1-2 | 59.4 | — |
| MathArena · Kangaroo 2025 11-12 | 87.5 | — |
| MathArena · Kangaroo 2025 3-4 | 50 | — |
| MathArena · Kangaroo 2025 5-6 | 58.3 | — |
| MathArena · Kangaroo 2025 7-8 | 86.7 | — |
| MathArena · Kangaroo 2025 9-10 | 91.7 | — |
| MathArena · USAMO 2026 | 47 | — |
| Vals Legal Research Bench | — | 38.5 |
Specs
| Spec | Opus 4.6 | Opus 4.7 |
|---|---|---|
| Input, per 1M | $5 | $5 |
| Output, per 1M | $25 | $25 |
| Speed, tok/s | 45 | 59 |
| Context | 1M | 1M |
| Weights | Closed | Closed |
| Released | Feb 2026 | Apr 2026 |
| See images | ✓ | ✓ |
| Read PDFs | Partial | Partial |
| Hear audio | — | — |
| Watch video | — | — |
| Call tools | ✓ | ✓ |
| Force a tool call | Partial | Partial |
| Return strict JSON | ✓ | ✓ |
| Search the web | ? | ? |
| Run code | ? | ? |
| Use a computer | ? | ? |
What breaks when you switch
Opus 4.6 → Opus 4.7
Nothing known.
Opus 4.7 → Opus 4.6
Nothing known.