Claude Opus 4.1 vs Claude Opus 4.5
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models
Opus 4.112.5
Opus 4.513.2
Arena Text · overall Arena Text · overall: Human pairwise preferences in the stated text vote pool
Opus 4.11,450
Opus 4.5 · high1,474
Arena Text · Chinese Arena Text · chinese: Human pairwise preferences in the stated text vote pool
Opus 4.11,427
Opus 4.51,469.8
Arena Text · English Arena Text · english: Human pairwise preferences in the stated text vote pool
Opus 4.11,426.5
Opus 4.51,460.6
Arena Text · French Arena Text · french: Human pairwise preferences in the stated text vote pool
Opus 4.11,427.4
Opus 4.51,470.8
Arena Text · German Arena Text · german: Human pairwise preferences in the stated text vote pool
Opus 4.11,404.9
Opus 4.51,449.4
Arena Text · instruction following Arena Text · instruction following: Human pairwise preferences in the stated text vote pool
Opus 4.11,433.6
Opus 4.51,472.5
Arena Text · Japanese Arena Text · japanese: Human pairwise preferences in the stated text vote pool
Opus 4.11,374.4
Opus 4.51,411.8
Arena Text · Korean Arena Text · korean: Human pairwise preferences in the stated text vote pool
Opus 4.11,364.1
Opus 4.51,423.6
Arena Text · non-English Arena Text · non english: Human pairwise preferences in the stated text vote pool
Opus 4.11,405.6
Opus 4.51,437.5
Arena Text · Polish Arena Text · polish: Human pairwise preferences in the stated text vote pool
Opus 4.11,413
Opus 4.51,438.7
Arena Text · Russian Arena Text · russian: Human pairwise preferences in the stated text vote pool
Opus 4.11,421.8
Opus 4.51,447.4
Arena Text · Spanish Arena Text · spanish: Human pairwise preferences in the stated text vote pool
Opus 4.11,448.2
Opus 4.51,452.5
SimpleQA Verified v1 SimpleQA Verified v1: Accuracy on short factual questions with verified answers, answered without web access
Opus 4.131
Opus 4.539
CritPt (physics) CritPt (physics): Research-level physics problems that need many careful reasoning steps
Opus 4.10
Opus 4.50.3
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 4.118.6
Opus 4.523.7
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 4.176
Opus 4.570.7
IFBench (instructions) IFBench (instructions): Checks whether the model follows precise formatting and content instructions
Opus 4.155.4
Opus 4.543
FrontierMath FrontierMath: Models solve expert-written advanced mathematics problems on the named tier and edition
Opus 4.15.9
Opus 4.520.7
Vals MedCode Vals MedCode: Medical coding accuracy on patient encounters
Opus 4.147.2
Opus 4.549.2
Vals MedScribe Vals MedScribe: Accuracy of clinical notes produced from consultation transcripts
Opus 4.173.9
Opus 4.585.3
AA Omniscience Index AA Omniscience Index: Scores factual knowledge while penalizing incorrect answers to encourage admitting uncertainty
Opus 4.1—
Opus 4.5-4.1
AA Omniscience · Accuracy AA Omniscience · Accuracy: Measures correct answers on the AA Omniscience factual knowledge test
Opus 4.1—
Opus 4.540.9
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 4.1—
Opus 4.576.2
EQ-Bench 3 EQ-Bench 3: Judge-rated responses to social and emotionally nuanced scenarios, scored through the source pairwise ranking
Opus 4.1—
Opus 4.51,545.7
Chess Puzzles Chess Puzzles: Models choose the best next move from a novel chess position
Opus 4.1—
Opus 4.55
FrontierMath Tier 4 FrontierMath Tier 4: Models solve expert-written advanced mathematics problems on the named tier and edition
Opus 4.1—
Opus 4.54.2
Vals LegalBench Vals LegalBench: Legal reasoning and legal-language understanding on LegalBench
Opus 4.1—
Opus 4.584.6
| Benchmark | Opus 4.1 | Opus 4.5 |
|---|---|---|
| Humanity's Last Exam | 12.5 | 13.2 |
| Arena Text · overall | 1,450 | 1,474 · high |
| Arena Text · Chinese | 1,427 | 1,469.8 |
| Arena Text · English | 1,426.5 | 1,460.6 |
| Arena Text · French | 1,427.4 | 1,470.8 |
| Arena Text · German | 1,404.9 | 1,449.4 |
| Arena Text · instruction following | 1,433.6 | 1,472.5 |
| Arena Text · Japanese | 1,374.4 | 1,411.8 |
| Arena Text · Korean | 1,364.1 | 1,423.6 |
| Arena Text · non-English | 1,405.6 | 1,437.5 |
| Arena Text · Polish | 1,413 | 1,438.7 |
| Arena Text · Russian | 1,421.8 | 1,447.4 |
| Arena Text · Spanish | 1,448.2 | 1,452.5 |
| SimpleQA Verified v1 | 31 | 39 |
| CritPt (physics) | 0 | 0.3 |
| AA Intelligence Index 4.3.2 | 18.6 | 23.7 |
| AA-LCR (long context) 1.1 | 76 | 70.7 |
| IFBench (instructions) | 55.4 | 43 |
| FrontierMath | 5.9 | 20.7 |
| Vals MedCode | 47.2 | 49.2 |
| Vals MedScribe | 73.9 | 85.3 |
| AA Omniscience Index | — | -4.1 |
| AA Omniscience · Accuracy | — | 40.9 |
| AA Omniscience · Hallucination Rate | — | 76.2 |
| EQ-Bench 3 | — | 1,545.7 |
| Chess Puzzles | — | 5 |
| FrontierMath Tier 4 | — | 4.2 |
| Vals LegalBench | — | 84.6 |
Specs
| Spec | Opus 4.1 | Opus 4.5 |
|---|---|---|
| Input, per 1M | $15 | $5 |
| Output, per 1M | $75 | $25 |
| Speed, tok/s | — | 51 |
| Context | 200K | 200K |
| Weights | Closed | Closed |
| Released | Aug 2025 | Nov 2025 |
| 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 4.1 → Opus 4.5
Nothing known.
Opus 4.5 → Opus 4.1