Claude Opus 5 vs GPT-6 Astra
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models
Opus 551.3
GPT-6 Astra52.7
Arena Text · overall Arena Text · overall: Human pairwise preferences in the stated text vote pool
Opus 5 · max1,490
GPT-6 Astra · max1,477
Arena Text · Chinese Arena Text · chinese: Human pairwise preferences in the stated text vote pool
Opus 5 · max1,573.4
GPT-6 Astra · max1,493
Arena Text · English Arena Text · english: Human pairwise preferences in the stated text vote pool
Opus 5 · max1,507.3
GPT-6 Astra · max1,447.9
Arena Text · French Arena Text · french: Human pairwise preferences in the stated text vote pool
Opus 5 · high1,519.5
GPT-6 Astra · max1,444.4
Arena Text · German Arena Text · german: Human pairwise preferences in the stated text vote pool
Opus 5 · max1,525.9
GPT-6 Astra · max1,457.1
Arena Text · instruction following Arena Text · instruction following: Human pairwise preferences in the stated text vote pool
Opus 5 · max1,517.8
GPT-6 Astra · max1,454.7
Arena Text · Japanese Arena Text · japanese: Human pairwise preferences in the stated text vote pool
Opus 5 · max1,515.4
GPT-6 Astra · max1,392.3
Arena Text · Korean Arena Text · korean: Human pairwise preferences in the stated text vote pool
Opus 5 · max1,520.7
GPT-6 Astra · max1,429.6
Arena Text · non-English Arena Text · non english: Human pairwise preferences in the stated text vote pool
Opus 5 · max1,500.5
GPT-6 Astra · max1,432.5
Arena Text · Polish Arena Text · polish: Human pairwise preferences in the stated text vote pool
Opus 5 · max1,519.9
GPT-6 Astra · max1,419.5
Arena Text · Russian Arena Text · russian: Human pairwise preferences in the stated text vote pool
Opus 5 · max1,507.2
GPT-6 Astra · max1,437
Arena Text · Spanish Arena Text · spanish: Human pairwise preferences in the stated text vote pool
Opus 5 · max1,519.5
GPT-6 Astra · max1,422.6
AA Omniscience Index AA Omniscience Index: Scores factual knowledge while penalizing incorrect answers to encourage admitting uncertainty
Opus 531
GPT-6 Astra42.2
SimpleQA Verified v1 SimpleQA Verified v1: Accuracy on short factual questions with verified answers, answered without web access
Opus 563.3
GPT-6 Astra75.8
CritPt (physics) CritPt (physics): Research-level physics problems that need many careful reasoning steps
Opus 526.9
GPT-6 Astra29.1
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 544.8
GPT-6 Astra49.6
AA Omniscience · Accuracy AA Omniscience · Accuracy: Measures correct answers on the AA Omniscience factual knowledge test
Opus 557.1
GPT-6 Astra60.6
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 560.7
GPT-6 Astra46.5
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 582
GPT-6 Astra79.7
GDP.pdf · AA all-pass GDP.pdf · AA all-pass: Models reason over professional PDF documents; an attempt passes only when every atomic criterion passes
Opus 520
GPT-6 Astra30.4
MLCR-AA MLCR-AA: Models synthesize long medical records; the headline credits concise answers judged both accurate and complete
Opus 556.1
GPT-6 Astra · max35
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 5 · max99
GPT-6 Astra98.8
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 5 · max97.5
GPT-6 Astra97.5
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 5 · max97.1
GPT-6 Astra96.7
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 5 · max90.4
GPT-6 Astra92.1
ARC-AGI-3 · semi-private ARC-AGI-3 · semi-private: Measures interactive game solving through relative human action efficiency
Opus 5 · high30.2
GPT-6 Astra38.6
Chess Puzzles Chess Puzzles: Models choose the best next move from a novel chess position
Opus 5 · max43
GPT-6 Astra · max74
FrontierMath Erdos FrontierMath Erdos: Models solve expert-written advanced mathematics problems on the named tier and edition
Opus 5 · max0
GPT-6 Astra · max2.9
FrontierMath Tier 4 (v2) FrontierMath Tier 4 (v2): Models solve expert-written advanced mathematics problems on the named tier and edition
Opus 5 · max73.2
GPT-6 Astra97.6
FrontierMath Tiers 1–3 (v2) FrontierMath Tiers 1–3 (v2): Models solve expert-written advanced mathematics problems on the named tier and edition
Opus 5 · max85.6
GPT-6 Astra · max93.7
Mystery Game Puzzles Mystery Game Puzzles: Models choose the best move in puzzle positions from an undisclosed game
Opus 5 · max59
GPT-6 Astra · max84
MathArena · ArXivMath · June 2026 MathArena · arxiv · 06/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 5 · max82.6
GPT-6 Astra · max94.4
MathArena · BrokenArXiv · June 2026 MathArena · arxiv_false · 06/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 5 · max90.7
GPT-6 Astra · max99.1
MathArena · ArXivLean · June 2026 MathArena · arxivlean · 06/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 5 · max32.6
GPT-6 Astra · max65.2
Vals MedCode Vals MedCode: Medical coding accuracy on patient encounters
Opus 563.6
GPT-6 Astra48.5
Vals MedScribe Vals MedScribe: Accuracy of clinical notes produced from consultation transcripts
Opus 591
GPT-6 Astra87.9
Vals Legal Research Bench Vals Legal Research Bench: Grounded legal research; a task passes only if all criteria pass
Opus 555.3
GPT-6 Astra39.4
Vals Science Index Vals Science Index: Scientific problem solving across the index’s component evaluations
Opus 547.9
GPT-6 Astra65.1
MathArena · ArXivMath · April 2026 MathArena · arxiv · 04/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 5—
GPT-6 Astra · max90.8
MathArena · ArXivMath · August 2026 MathArena · arxiv · 08/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 5—
GPT-6 Astra · max88.6
MathArena · ArXivMath · May 2026 MathArena · arxiv · 05/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 5—
GPT-6 Astra · max95
MathArena · BrokenArXiv · April 2026 MathArena · arxiv_false · 04/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 5—
GPT-6 Astra · max97.5
MathArena · BrokenArXiv · August 2026 MathArena · arxiv_false · 08/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 5—
GPT-6 Astra · max81.9
MathArena · BrokenArXiv · May 2026 MathArena · arxiv_false · 05/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 5—
GPT-6 Astra · max91
MathArena · ArXivLean · September 2026 MathArena · arxivlean · 09/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 5—
GPT-6 Astra · max37.7
Vals LegalBench Vals LegalBench: Legal reasoning and legal-language understanding on LegalBench
Opus 587
GPT-6 Astra—
| Benchmark | Opus 5 | GPT-6 Astra |
|---|---|---|
| Humanity's Last Exam | 51.3 | 52.7 |
| Arena Text · overall | 1,490 · max | 1,477 · max |
| Arena Text · Chinese | 1,573.4 · max | 1,493 · max |
| Arena Text · English | 1,507.3 · max | 1,447.9 · max |
| Arena Text · French | 1,519.5 · high | 1,444.4 · max |
| Arena Text · German | 1,525.9 · max | 1,457.1 · max |
| Arena Text · instruction following | 1,517.8 · max | 1,454.7 · max |
| Arena Text · Japanese | 1,515.4 · max | 1,392.3 · max |
| Arena Text · Korean | 1,520.7 · max | 1,429.6 · max |
| Arena Text · non-English | 1,500.5 · max | 1,432.5 · max |
| Arena Text · Polish | 1,519.9 · max | 1,419.5 · max |
| Arena Text · Russian | 1,507.2 · max | 1,437 · max |
| Arena Text · Spanish | 1,519.5 · max | 1,422.6 · max |
| AA Omniscience Index | 31 | 42.2 |
| SimpleQA Verified v1 | 63.3 | 75.8 |
| CritPt (physics) | 26.9 | 29.1 |
| AA Intelligence Index 4.3.2 | 44.8 | 49.6 |
| AA Omniscience · Accuracy | 57.1 | 60.6 |
| AA Omniscience · Hallucination Rate | 60.7 | 46.5 |
| AA-LCR (long context) 1.1 | 82 | 79.7 |
| GDP.pdf · AA all-pass | 20 | 30.4 |
| MLCR-AA | 56.1 | 35 · max |
| ARC-AGI-1 · public evaluation | 99 · max | 98.8 |
| ARC-AGI-1 · semi-private | 97.5 · max | 97.5 |
| ARC-AGI-2 · public evaluation | 97.1 · max | 96.7 |
| ARC-AGI-2 · semi-private | 90.4 · max | 92.1 |
| ARC-AGI-3 · semi-private | 30.2 · high | 38.6 |
| Chess Puzzles | 43 · max | 74 · max |
| FrontierMath Erdos | 0 · max | 2.9 · max |
| FrontierMath Tier 4 (v2) | 73.2 · max | 97.6 |
| FrontierMath Tiers 1–3 (v2) | 85.6 · max | 93.7 · max |
| Mystery Game Puzzles | 59 · max | 84 · max |
| MathArena · ArXivMath · June 2026 | 82.6 · max | 94.4 · max |
| MathArena · BrokenArXiv · June 2026 | 90.7 · max | 99.1 · max |
| MathArena · ArXivLean · June 2026 | 32.6 · max | 65.2 · max |
| Vals MedCode | 63.6 | 48.5 |
| Vals MedScribe | 91 | 87.9 |
| Vals Legal Research Bench | 55.3 | 39.4 |
| Vals Science Index | 47.9 | 65.1 |
| MathArena · ArXivMath · April 2026 | — | 90.8 · max |
| MathArena · ArXivMath · August 2026 | — | 88.6 · max |
| MathArena · ArXivMath · May 2026 | — | 95 · max |
| MathArena · BrokenArXiv · April 2026 | — | 97.5 · max |
| MathArena · BrokenArXiv · August 2026 | — | 81.9 · max |
| MathArena · BrokenArXiv · May 2026 | — | 91 · max |
| MathArena · ArXivLean · September 2026 | — | 37.7 · max |
| Vals LegalBench | 87 | — |
Specs
| Spec | Opus 5 | GPT-6 Astra |
|---|---|---|
| Input, per 1M | $5 | $10 |
| Output, per 1M | $25 | $50 |
| Cost per task? | $2.19 | $1.54 |
| Speed, tok/s | 55 | 59 |
| Context | 1M | 1.05M |
| Weights | Closed | Closed |
| Released | Jul 2026 | Sep 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 5 → GPT-6 Astra
Monitoring can return HTTP 403›
Misalignment monitoring can return HTTP 403 before streaming or errors after output; match misalignment_policy_violation and handle stream errors.
No none reasoning effort›
The none reasoning effort is unsupported on GPT-6 Astra and GPT-6.1 Sol; use low instead.
Remove unsupported sampling parameters›
When reasoning effort is not none, remove temperature, top_p and top_logprobs, plus Chat Completions logprobs or Responses output logprobs inclusion.
GPT-6 Astra → Opus 5