Model comparison
Claude Opus 4.8 vs Gemini 3 Pro
Claude Opus 4.8 is the stronger model overall, scoring 60.7 to 54.8 on the Noometry Index.
Last verified . 45 shared benchmarks.
GoogleSummary
- They share 45 benchmarks with published results for both. Claude Opus 4.8 scores higher in 7 categories and Gemini 3 Pro in 3 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where Claude Opus 4.8 leads 78.4 to 49.9.
- The biggest single-benchmark swing is ProofBench: 69% for Claude Opus 4.8 and 20% for Gemini 3 Pro.
Side by side
| Claude Opus 4.8 | Gemini 3 Pro | |
|---|---|---|
| Provider | Anthropic | |
| Noometry Index | 60.7 | 54.8 |
| Released | 2026-05-28 | 2025-11-18 |
| Weights | Proprietary | Proprietary |
| Context window | 1M | — |
| Max output | 128K | — |
| Input $ / M tokens | $5 | — |
| Output $ / M tokens | $25 | — |
| Results tracked | 65 | 67 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Opus 4.8 leads
Claude Opus 4.8: 59.9 (#12), Gemini 3 Pro: 51.6 (#39)
| Benchmark | Claude Opus 4.8 | Gemini 3 Pro |
|---|---|---|
| LMArena WebDev | 1556 | 1440 |
| GSO | 47.1% | 18.6% |
| WeirdML | 82.9% | 69.9% |
| LMArena Coding | 1490 | 1481 |
| ALE-Bench | 1,564 | 1,177 |
| SWE-bench Verified | — | 72.9% |
| DeepSWE | 59% | — |
| FrontierCode | 46.5% | — |
| SWE-bench Verified (bash only) | — | 74.2% |
| SWE-bench Multilingual | — | 68.7% |
| SciCode | 53.5% | — |
| AlgoTune | — | 1.83 |
Agentic & Tool Use Claude Opus 4.8 leads
Claude Opus 4.8: 47.6 (#11), Gemini 3 Pro: 40.6 (#23)
| Benchmark | Claude Opus 4.8 | Gemini 3 Pro |
|---|---|---|
| Remote Labor Index | 8.3% | 1.3% |
| τ²-bench Banking | 39.7% | 18% |
| DeepResearch Bench | 50.2% | 46.3% |
| LMArena Search | 1204 | 1207 |
| Vending-Bench 2 | 5,787 | 5,478 |
| Terminal-Bench | — | 69.4% |
| APEX-Agents | 48.9% | — |
| Berkeley Function Calling Leaderboard | — | 72.5% |
| OSWorld 2.0 | 20.6% | — |
| GDPval | — | 40.3% |
| τ²-bench Airline | — | 80.5% |
| τ²-bench Retail | — | 75.9% |
| τ²-bench Telecom | — | 91% |
| PostTrainBench | 33.8% | — |
| BALROG | — | 58.1% |
| GBAEval | 70.9% | — |
| GDP.pdf | 24% | — |
| METR Time Horizons | — | 71% |
Reasoning Claude Opus 4.8 leads
Claude Opus 4.8: 64.7 (#16), Gemini 3 Pro: 52.5 (#31)
| Benchmark | Claude Opus 4.8 | Gemini 3 Pro |
|---|---|---|
| ARC-AGI-2 | 72.1% | 31.1% |
| SimpleBench | 64.8% | 76.4% |
| Kagi LLM Benchmark | 88.8% | 80.1% |
| NYT Connections (extended) | 91.1% | 94.4% |
| ARC-AGI-1 | 92.5% | 75% |
| CritPt | 20.9% | 6.9% |
| Chess Puzzles | 34% | 31% |
| EnigmaEval | 23.5% | 18.2% |
| LMArena Hard Prompts | 1482 | 1480 |
| Epoch Capabilities Index | 158.21 | 152.92 |
| ForecastBench | 59.9 | 61.2 |
| EBR-Bench | 28.6% | — |
| Mystery Game Puzzles | 36% | — |
| DTBench | 94.9% | — |
| LMCA | 57.5% | — |
| Surface Evolver Bench | 87.5% | — |
| Bench to the Future 3 | 0.14 | — |
Math Claude Opus 4.8 leads
Claude Opus 4.8: 78.4 (#13), Gemini 3 Pro: 49.9 (#59)
| Benchmark | Claude Opus 4.8 | Gemini 3 Pro |
|---|---|---|
| MathArena Final-Answer Competitions | 91.8% | 67% |
| OTIS Mock AIME 2024-2025 | 98.3% | 91.4% |
| ProofBench | 69% | 20% |
| LMArena Math | 1487 | 1476 |
| FrontierMath (Feb 2025 set) | 47.2% | 37.6% |
| FrontierMath Tier 4 (v1) | 31.3% | 18.8% |
| FrontierMath (Tiers 1-3) | 80% | — |
| FrontierMath Tier 4 | 56.1% | — |
| Omni-MATH | — | 55.5% |
Knowledge Gemini 3 Pro leads
Claude Opus 4.8: 61.3 (#29), Gemini 3 Pro: 64.4 (#16)
| Benchmark | Claude Opus 4.8 | Gemini 3 Pro |
|---|---|---|
| GPQA Diamond | 91% | 92.6% |
| LMArena Expert | 1502 | 1475 |
| Humanity's Last Exam | — | 37.5% |
| SimpleQA Verified | 53% | — |
| MMLU-Pro | — | 90.3% |
| Vectara Hallucination Rate | — | 13.6% |
| GPQA (HELM) | — | 80.3% |
Multimodal Gemini 3 Pro leads
Claude Opus 4.8: 42.9 (#26), Gemini 3 Pro: 57.6 (#2)
| Benchmark | Claude Opus 4.8 | Gemini 3 Pro |
|---|---|---|
| LMArena Vision | 1294 | 1305 |
| LMArena Document | 1475 | 1434 |
| GeoBench | — | 84% |
| VPCT | — | 91% |
| Blueprint-Bench 2 | 14.5% | — |
| Furniture Assembly | 42.5% | — |
Multilingual Gemini 3 Pro leads
Claude Opus 4.8: 55.2 (#33), Gemini 3 Pro: 56.9 (#16)
| Benchmark | Claude Opus 4.8 | Gemini 3 Pro |
|---|---|---|
| LMArena Non-English | 1450 | 1474 |
| LMArena Chinese | 1507 | 1523 |
| LMArena French | 1481 | 1492 |
| LMArena German | 1472 | 1515 |
| LMArena Japanese | 1440 | 1510 |
| LMArena Korean | 1432 | 1448 |
| LMArena Russian | 1474 | 1493 |
| LMArena Spanish | 1466 | 1470 |
Instruction Following Claude Opus 4.8 leads
Claude Opus 4.8: 77.4 (#24), Gemini 3 Pro: 76.3 (#45)
| Benchmark | Claude Opus 4.8 | Gemini 3 Pro |
|---|---|---|
| LMArena Instruction Following | 1476 | 1458 |
| IFEval | — | 87.7% |
Long Context Claude Opus 4.8 leads
Claude Opus 4.8: 45.4 (#35), Gemini 3 Pro: 44.0 (#79)
| Benchmark | Claude Opus 4.8 | Gemini 3 Pro |
|---|---|---|
| LMArena Longer Query | 1483 | 1471 |
| CL-bench | — | 15.8% |
Writing & Preference Claude Opus 4.8 leads
Claude Opus 4.8: 72.0 (#16), Gemini 3 Pro: 66.4 (#35)
| Benchmark | Claude Opus 4.8 | Gemini 3 Pro |
|---|---|---|
| LMArena Text | 1461 | 1479 |
| LMArena Creative Writing | 1454 | 1482 |
| EQ-Bench Creative Writing | 1840 | 1525 |
| LMArena Multi-Turn | 1476 | 1484 |
| WildBench | — | 85.9% |
| EQ-Bench 4 | 1281 | — |
Frequently asked questions
Is Claude Opus 4.8 better than Gemini 3 Pro?
Claude Opus 4.8 is the stronger model overall, scoring 60.7 to 54.8 on the Noometry Index.
Is Claude Opus 4.8 or Gemini 3 Pro better for coding?
Claude Opus 4.8 scores higher on coding benchmarks: 59.9 versus 51.6 in the Noometry coding category.
How many benchmarks do Claude Opus 4.8 and Gemini 3 Pro share?
45 benchmarks have published results for both models. Claude Opus 4.8 has 65 scored results on Noometry and Gemini 3 Pro has 67.