Claude Opus 4.7 vs Gemini 3.8 Flash
Different efforts: high · medium · Compare all at high
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models
Opus 4.733.3
Gemini 3.8 Flash42.1
Arena Text · overall Arena Text · overall: Human pairwise preferences in the stated text vote pool
Opus 4.71,501
Gemini 3.8 Flash · high1,496
Arena Text · Chinese Arena Text · chinese: Human pairwise preferences in the stated text vote pool
Opus 4.71,531.8
Gemini 3.8 Flash · high1,550.8
Arena Text · English Arena Text · english: Human pairwise preferences in the stated text vote pool
Opus 4.71,496.3
Gemini 3.8 Flash · high1,499
Arena Text · French Arena Text · french: Human pairwise preferences in the stated text vote pool
Opus 4.71,503
Gemini 3.8 Flash · high1,491.7
Arena Text · German Arena Text · german: Human pairwise preferences in the stated text vote pool
Opus 4.71,495.2
Gemini 3.8 Flash · high1,497.6
Arena Text · instruction following Arena Text · instruction following: Human pairwise preferences in the stated text vote pool
Opus 4.71,498
Gemini 3.8 Flash · high1,491
Arena Text · Japanese Arena Text · japanese: Human pairwise preferences in the stated text vote pool
Opus 4.71,473.9
Gemini 3.8 Flash · high1,506.3
Arena Text · Korean Arena Text · korean: Human pairwise preferences in the stated text vote pool
Opus 4.71,450.9
Gemini 3.8 Flash · high1,461.6
Arena Text · non-English Arena Text · non english: Human pairwise preferences in the stated text vote pool
Opus 4.71,480.5
Gemini 3.8 Flash · high1,491.1
Arena Text · Polish Arena Text · polish: Human pairwise preferences in the stated text vote pool
Opus 4.71,479.5
Gemini 3.8 Flash · high1,498.5
Arena Text · Russian Arena Text · russian: Human pairwise preferences in the stated text vote pool
Opus 4.71,494.4
Gemini 3.8 Flash · high1,514.4
Arena Text · Spanish Arena Text · spanish: Human pairwise preferences in the stated text vote pool
Opus 4.71,494.6
Gemini 3.8 Flash · high1,485.7
AA Omniscience Index AA Omniscience Index: Scores factual knowledge while penalizing incorrect answers to encourage admitting uncertainty
Opus 4.714.8
Gemini 3.8 Flash28.6
SimpleQA Verified v1 SimpleQA Verified v1: Accuracy on short factual questions with verified answers, answered without web access
Opus 4.7 · xhigh51.7
Gemini 3.8 Flash74.6
CritPt (physics) CritPt (physics): Research-level physics problems that need many careful reasoning steps
Opus 4.75.1
Gemini 3.8 Flash12.3
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 4.730.9
Gemini 3.8 Flash39.8
AA Omniscience · Accuracy AA Omniscience · Accuracy: Measures correct answers on the AA Omniscience factual knowledge test
Opus 4.744.7
Gemini 3.8 Flash53
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 4.754.1
Gemini 3.8 Flash51.9
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 4.775.7
Gemini 3.8 Flash84
MLCR-AA MLCR-AA: Models synthesize long medical records; the headline credits concise answers judged both accurate and complete
Opus 4.728.9
Gemini 3.8 Flash · high21.7
Chess Puzzles Chess Puzzles: Models choose the best next move from a novel chess position
Opus 4.7 · max41
Gemini 3.8 Flash · high58
FrontierMath Tier 4 (v2) FrontierMath Tier 4 (v2): Models solve expert-written advanced mathematics problems on the named tier and edition
Opus 4.7 · max31.7
Gemini 3.8 Flash · high22
FrontierMath Tiers 1–3 (v2) FrontierMath Tiers 1–3 (v2): Models solve expert-written advanced mathematics problems on the named tier and edition
Opus 4.7 · max70.2
Gemini 3.8 Flash · high68.4
Mystery Game Puzzles Mystery Game Puzzles: Models choose the best move in puzzle positions from an undisclosed game
Opus 4.7 · max28
Gemini 3.8 Flash · high47
MathArena · ArXivMath · April 2026 MathArena · arxiv · 04/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.7 · xhigh57.5
Gemini 3.8 Flash64.2
MathArena · BrokenArXiv · April 2026 MathArena · arxiv_false · 04/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.7 · xhigh4.1
Gemini 3.8 Flash33.6
Vals LegalBench Vals LegalBench: Legal reasoning and legal-language understanding on LegalBench
Opus 4.785.3
Gemini 3.8 Flash87
Vals MedCode Vals MedCode: Medical coding accuracy on patient encounters
Opus 4.754.9
Gemini 3.8 Flash48.1
Vals MedScribe Vals MedScribe: Accuracy of clinical notes produced from consultation transcripts
Opus 4.783
Gemini 3.8 Flash84.5
Vals Legal Research Bench Vals Legal Research Bench: Grounded legal research; a task passes only if all criteria pass
Opus 4.738.5
Gemini 3.8 Flash38.9
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 4.7—
Gemini 3.8 Flash22.8
IFBench (instructions) IFBench (instructions): Checks whether the model follows precise formatting and content instructions
Opus 4.743.6
Gemini 3.8 Flash—
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 4.7—
Gemini 3.8 Flash97.8
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 4.7—
Gemini 3.8 Flash97.5
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 4.7—
Gemini 3.8 Flash87.5
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 4.7—
Gemini 3.8 Flash82.9
ARC-AGI-3 · semi-private ARC-AGI-3 · semi-private: Measures interactive game solving through relative human action efficiency
Opus 4.7—
Gemini 3.8 Flash4
EQ-Bench 3 EQ-Bench 3: Judge-rated responses to social and emotionally nuanced scenarios, scored through the source pairwise ranking
Opus 4.71,884.3
Gemini 3.8 Flash—
FrontierMath FrontierMath: Models solve expert-written advanced mathematics problems on the named tier and edition
Opus 4.7 · xhigh43.8
Gemini 3.8 Flash—
FrontierMath Erdos FrontierMath Erdos: Models solve expert-written advanced mathematics problems on the named tier and edition
Opus 4.7—
Gemini 3.8 Flash · high0
FrontierMath Tier 4 FrontierMath Tier 4: Models solve expert-written advanced mathematics problems on the named tier and edition
Opus 4.7 · xhigh22.9
Gemini 3.8 Flash—
MathArena · Apex MathArena · apex · Apex: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.7 · xhigh40.6
Gemini 3.8 Flash—
MathArena · Apex Shortlist MathArena · apex · Apex Shortlist: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.7 · xhigh63.8
Gemini 3.8 Flash—
MathArena · ArXivMath · August 2026 MathArena · arxiv · 08/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.7—
Gemini 3.8 Flash40.4
MathArena · ArXivMath · February 2026 MathArena · arxiv · 02/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.7 · xhigh40.6
Gemini 3.8 Flash—
MathArena · ArXivMath · January 2026 MathArena · arxiv · 01/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.7 · xhigh52.2
Gemini 3.8 Flash—
MathArena · ArXivMath · June 2026 MathArena · arxiv · 06/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.7—
Gemini 3.8 Flash67.4
MathArena · ArXivMath · March 2026 MathArena · arxiv · 03/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.7 · xhigh50.8
Gemini 3.8 Flash—
MathArena · ArXivMath · May 2026 MathArena · arxiv · 05/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.7—
Gemini 3.8 Flash58.3
MathArena · ArXivMath · September 2026 MathArena · arxiv · 09/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.7—
Gemini 3.8 Flash50
MathArena · BrokenArXiv · August 2026 MathArena · arxiv_false · 08/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.7—
Gemini 3.8 Flash27.4
MathArena · BrokenArXiv · February 2026 MathArena · arxiv_false · 02/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.7 · xhigh4
Gemini 3.8 Flash—
MathArena · BrokenArXiv · June 2026 MathArena · arxiv_false · 06/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.7—
Gemini 3.8 Flash20.8
MathArena · BrokenArXiv · March 2026 MathArena · arxiv_false · 03/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.7 · xhigh5.8
Gemini 3.8 Flash—
MathArena · BrokenArXiv · May 2026 MathArena · arxiv_false · 05/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.7—
Gemini 3.8 Flash19
MathArena · BrokenArXiv · September 2026 MathArena · arxiv_false · 09/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.7—
Gemini 3.8 Flash26.7
MathArena · ArXivLean · June 2026 MathArena · arxivlean · 06/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.7—
Gemini 3.8 Flash19.6
MathArena · ArXivLean · March 2026 MathArena · arxivlean · 03/2026: Answer accuracy on the named mathematics competition under the recorded effort
Opus 4.714.6
Gemini 3.8 Flash—
Vals Science Index Vals Science Index: Scientific problem solving across the index’s component evaluations
Opus 4.7—
Gemini 3.8 Flash35.8
| Benchmark | Opus 4.7 | Gemini 3.8 Flash |
|---|---|---|
| Humanity's Last Exam | 33.3 | 42.1 |
| Arena Text · overall | 1,501 | 1,496 · high |
| Arena Text · Chinese | 1,531.8 | 1,550.8 · high |
| Arena Text · English | 1,496.3 | 1,499 · high |
| Arena Text · French | 1,503 | 1,491.7 · high |
| Arena Text · German | 1,495.2 | 1,497.6 · high |
| Arena Text · instruction following | 1,498 | 1,491 · high |
| Arena Text · Japanese | 1,473.9 | 1,506.3 · high |
| Arena Text · Korean | 1,450.9 | 1,461.6 · high |
| Arena Text · non-English | 1,480.5 | 1,491.1 · high |
| Arena Text · Polish | 1,479.5 | 1,498.5 · high |
| Arena Text · Russian | 1,494.4 | 1,514.4 · high |
| Arena Text · Spanish | 1,494.6 | 1,485.7 · high |
| AA Omniscience Index | 14.8 | 28.6 |
| SimpleQA Verified v1 | 51.7 · xhigh | 74.6 |
| CritPt (physics) | 5.1 | 12.3 |
| AA Intelligence Index 4.3.2 | 30.9 | 39.8 |
| AA Omniscience · Accuracy | 44.7 | 53 |
| AA Omniscience · Hallucination Rate | 54.1 | 51.9 |
| AA-LCR (long context) 1.1 | 75.7 | 84 |
| MLCR-AA | 28.9 | 21.7 · high |
| Chess Puzzles | 41 · max | 58 · high |
| FrontierMath Tier 4 (v2) | 31.7 · max | 22 · high |
| FrontierMath Tiers 1–3 (v2) | 70.2 · max | 68.4 · high |
| Mystery Game Puzzles | 28 · max | 47 · high |
| MathArena · ArXivMath · April 2026 | 57.5 · xhigh | 64.2 |
| MathArena · BrokenArXiv · April 2026 | 4.1 · xhigh | 33.6 |
| Vals LegalBench | 85.3 | 87 |
| Vals MedCode | 54.9 | 48.1 |
| Vals MedScribe | 83 | 84.5 |
| Vals Legal Research Bench | 38.5 | 38.9 |
| GDP.pdf · AA all-pass | — | 22.8 |
| IFBench (instructions) | 43.6 | — |
| ARC-AGI-1 · public evaluation | — | 97.8 |
| ARC-AGI-1 · semi-private | — | 97.5 |
| ARC-AGI-2 · public evaluation | — | 87.5 |
| ARC-AGI-2 · semi-private | — | 82.9 |
| ARC-AGI-3 · semi-private | — | 4 |
| EQ-Bench 3 | 1,884.3 | — |
| FrontierMath | 43.8 · xhigh | — |
| FrontierMath Erdos | — | 0 · high |
| FrontierMath Tier 4 | 22.9 · xhigh | — |
| MathArena · Apex | 40.6 · xhigh | — |
| MathArena · Apex Shortlist | 63.8 · xhigh | — |
| MathArena · ArXivMath · August 2026 | — | 40.4 |
| MathArena · ArXivMath · February 2026 | 40.6 · xhigh | — |
| MathArena · ArXivMath · January 2026 | 52.2 · xhigh | — |
| MathArena · ArXivMath · June 2026 | — | 67.4 |
| MathArena · ArXivMath · March 2026 | 50.8 · xhigh | — |
| MathArena · ArXivMath · May 2026 | — | 58.3 |
| MathArena · ArXivMath · September 2026 | — | 50 |
| MathArena · BrokenArXiv · August 2026 | — | 27.4 |
| MathArena · BrokenArXiv · February 2026 | 4 · xhigh | — |
| MathArena · BrokenArXiv · June 2026 | — | 20.8 |
| MathArena · BrokenArXiv · March 2026 | 5.8 · xhigh | — |
| MathArena · BrokenArXiv · May 2026 | — | 19 |
| MathArena · BrokenArXiv · September 2026 | — | 26.7 |
| MathArena · ArXivLean · June 2026 | — | 19.6 |
| MathArena · ArXivLean · March 2026 | 14.6 | — |
| Vals Science Index | — | 35.8 |
Specs
| Spec | Opus 4.7 | Gemini 3.8 Flash |
|---|---|---|
| Input, per 1M | $5 | !$0.75 |
| Output, per 1M | $25 | !$3.75 |
| Cost per task? | — | $0.93 |
| Speed, tok/s | 59 | — |
| Context | 1M | 1M |
| Weights | Closed | Closed |
| Released | Apr 2026 | Sep 2026 |
| See images | ✓ | ✓ |
| Read PDFs | 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 4.7 → Gemini 3.8 Flash
Sampling parameters are ignored›
temperature, top_p and top_k are ignored now and will cause HTTP 400 in future generations; remove them from requests.
No minimal thinking level›
The minimal thinking level returns an error; avoid minimal.
Gemini 3.8 Flash → Opus 4.7
Nothing known.