Model comparison
Claude Sonnet 4.5 vs Gemini 3 Pro
Gemini 3 Pro is the stronger model overall, scoring 54.8 to 44.1 on the Noometry Index.
Last verified . 62 shared benchmarks.
GoogleSummary
- They share 62 benchmarks with published results for both. Claude Sonnet 4.5 scores higher in 2 categories and Gemini 3 Pro in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Gemini 3 Pro leads 52.5 to 26.9.
- The biggest single-benchmark swing is NYT Connections (extended): 37.3% for Claude Sonnet 4.5 and 94.4% for Gemini 3 Pro.
Side by side
| Claude Sonnet 4.5 | Gemini 3 Pro | |
|---|---|---|
| Provider | Anthropic | |
| Noometry Index | 44.1 | 54.8 |
| Released | 2025-09-29 | 2025-11-18 |
| Weights | Proprietary | Proprietary |
| Context window | 200K | — |
| Max output | 64K | — |
| Input $ / M tokens | $3 | — |
| Output $ / M tokens | $15 | — |
| Results tracked | 73 | 67 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemini 3 Pro leads
Claude Sonnet 4.5: 47.3 (#61), Gemini 3 Pro: 51.6 (#39)
| Benchmark | Claude Sonnet 4.5 | Gemini 3 Pro |
|---|---|---|
| SWE-bench Verified | 71.3% | 72.9% |
| SWE-bench Verified (bash only) | 71.4% | 74.2% |
| LMArena WebDev | 1393 | 1440 |
| SWE-bench Multilingual | 67% | 68.7% |
| GSO | 14.7% | 18.6% |
| WeirdML | 47.7% | 69.9% |
| LMArena Coding | 1489 | 1481 |
| ALE-Bench | 796.15 | 1,177 |
| AlgoTune | 1.52 | 1.83 |
| SciCode | 44.7% | — |
Agentic & Tool Use Gemini 3 Pro leads
Claude Sonnet 4.5: 38.3 (#32), Gemini 3 Pro: 40.6 (#23)
| Benchmark | Claude Sonnet 4.5 | Gemini 3 Pro |
|---|---|---|
| Terminal-Bench | 46.5% | 69.4% |
| Berkeley Function Calling Leaderboard | 73.2% | 72.5% |
| GDPval | 42.5% | 40.3% |
| Remote Labor Index | 2.1% | 1.3% |
| τ²-bench Airline | 72% | 80.5% |
| τ²-bench Banking | 25.3% | 18% |
| τ²-bench Retail | 72.4% | 75.9% |
| τ²-bench Telecom | 84.9% | 91% |
| DeepResearch Bench | 52.6% | 46.3% |
| LMArena Search | 1159 | 1207 |
| METR Time Horizons | 67.4% | 71% |
| Vending-Bench 2 | 3,839 | 5,478 |
| Cybench | 60% | — |
| OSWorld | 62.9% | — |
| BALROG | — | 58.1% |
Reasoning Gemini 3 Pro leads
Claude Sonnet 4.5: 26.9 (#125), Gemini 3 Pro: 52.5 (#31)
| Benchmark | Claude Sonnet 4.5 | Gemini 3 Pro |
|---|---|---|
| ARC-AGI-2 | 13.6% | 31.1% |
| SimpleBench | 54.3% | 76.4% |
| Kagi LLM Benchmark | 57.9% | 80.1% |
| NYT Connections (extended) | 37.3% | 94.4% |
| ARC-AGI-1 | 63.7% | 75% |
| CritPt | 1.1% | 6.9% |
| Chess Puzzles | 12% | 31% |
| EnigmaEval | 6% | 18.2% |
| LMArena Hard Prompts | 1462 | 1480 |
| Epoch Capabilities Index | 146.84 | 152.92 |
| ForecastBench | 61.9 | 61.2 |
| EBR-Bench | 2.4% | — |
| Mystery Game Puzzles | 17% | — |
| DTBench | 83.2% | — |
| LMCA | 38.8% | — |
Math Gemini 3 Pro leads
Claude Sonnet 4.5: 32.3 (#216), Gemini 3 Pro: 49.9 (#59)
| Benchmark | Claude Sonnet 4.5 | Gemini 3 Pro |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 77.8% | 91.4% |
| ProofBench | 19% | 20% |
| Omni-MATH | 55.3% | 55.5% |
| LMArena Math | 1449 | 1476 |
| FrontierMath (Feb 2025 set) | 15.2% | 37.6% |
| FrontierMath Tier 4 (v1) | 4.2% | 18.8% |
| FrontierMath (Tiers 1-3) | 23.9% | — |
| FrontierMath Tier 4 | 2.4% | — |
| MathArena Final-Answer Competitions | — | 67% |
| MATH Level 5 | 97.7% | — |
Knowledge Gemini 3 Pro leads
Claude Sonnet 4.5: 48.4 (#76), Gemini 3 Pro: 64.4 (#16)
| Benchmark | Claude Sonnet 4.5 | Gemini 3 Pro |
|---|---|---|
| GPQA Diamond | 82.3% | 92.6% |
| Humanity's Last Exam | 13.7% | 37.5% |
| MMLU-Pro | 86.9% | 90.3% |
| Vectara Hallucination Rate | 12% | 13.6% |
| GPQA (HELM) | 68.6% | 80.3% |
| LMArena Expert | 1482 | 1475 |
| SimpleQA Verified | 30.7% | — |
Multimodal Gemini 3 Pro leads
Claude Sonnet 4.5: 34.8 (#89), Gemini 3 Pro: 57.6 (#2)
| Benchmark | Claude Sonnet 4.5 | Gemini 3 Pro |
|---|---|---|
| VPCT | 39.8% | 91% |
| LMArena Document | 1450 | 1434 |
| LMArena Vision | — | 1305 |
| GeoBench | — | 84% |
Multilingual Gemini 3 Pro leads
Claude Sonnet 4.5: 53.4 (#69), Gemini 3 Pro: 56.9 (#16)
| Benchmark | Claude Sonnet 4.5 | Gemini 3 Pro |
|---|---|---|
| LMArena Non-English | 1425 | 1474 |
| LMArena Chinese | 1459 | 1523 |
| LMArena French | 1458 | 1492 |
| LMArena German | 1427 | 1515 |
| LMArena Japanese | 1390 | 1510 |
| LMArena Korean | 1403 | 1448 |
| LMArena Russian | 1437 | 1493 |
| LMArena Spanish | 1457 | 1470 |
Instruction Following Gemini 3 Pro leads
Claude Sonnet 4.5: 75.0 (#78), Gemini 3 Pro: 76.3 (#45)
| Benchmark | Claude Sonnet 4.5 | Gemini 3 Pro |
|---|---|---|
| IFEval | 85% | 87.7% |
| LMArena Instruction Following | 1459 | 1458 |
Long Context Claude Sonnet 4.5 leads
Claude Sonnet 4.5: 45.2 (#46), Gemini 3 Pro: 44.0 (#79)
| Benchmark | Claude Sonnet 4.5 | Gemini 3 Pro |
|---|---|---|
| LMArena Longer Query | 1476 | 1471 |
| CL-bench | — | 15.8% |
Writing & Preference Too close to call
Claude Sonnet 4.5: 66.5 (#34), Gemini 3 Pro: 66.4 (#35)
| Benchmark | Claude Sonnet 4.5 | Gemini 3 Pro |
|---|---|---|
| LMArena Text | 1439 | 1479 |
| LMArena Creative Writing | 1442 | 1482 |
| EQ-Bench Creative Writing | 1678 | 1525 |
| WildBench | 85.4% | 85.9% |
| LMArena Multi-Turn | 1465 | 1484 |
Frequently asked questions
Is Claude Sonnet 4.5 better than Gemini 3 Pro?
Gemini 3 Pro is the stronger model overall, scoring 54.8 to 44.1 on the Noometry Index.
Is Claude Sonnet 4.5 or Gemini 3 Pro better for coding?
Gemini 3 Pro scores higher on coding benchmarks: 51.6 versus 47.3 in the Noometry coding category.
How many benchmarks do Claude Sonnet 4.5 and Gemini 3 Pro share?
62 benchmarks have published results for both models. Claude Sonnet 4.5 has 73 scored results on Noometry and Gemini 3 Pro has 67.