Model comparison
Gemini 3 Pro vs Kimi K2.5
Gemini 3 Pro is the stronger model overall, scoring 54.7 to 47.8 on the Noometry Index.
Last verified . 45 shared benchmarks.
GoogleSummary
- They share 45 benchmarks with published results for both. Gemini 3 Pro scores higher in 8 categories and Kimi K2.5 in 2 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Gemini 3 Pro leads 52.5 to 31.2.
- The biggest single-benchmark swing is SimpleBench: 76.4% for Gemini 3 Pro and 46.8% for Kimi K2.5.
- Kimi K2.5 has downloadable open weights; the other is API-only.
Side by side
| Gemini 3 Pro | Kimi K2.5 | |
|---|---|---|
| Provider | Moonshot AI | |
| Noometry Index | 54.7 | 47.8 |
| Released | 2025-11-18 | 2026-01-27 |
| Weights | Proprietary | Open |
| Context window | — | 262K |
| Max output | — | 262K |
| Input $ / M tokens | — | $0.45 |
| Output $ / M tokens | — | $2.25 |
| Results tracked | 67 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemini 3 Pro leads
Gemini 3 Pro: 52.1 (#33), Kimi K2.5: 48.7 (#48)
| Benchmark | Gemini 3 Pro | Kimi K2.5 |
|---|---|---|
| SWE-bench Verified | 72.9% | 73.8% |
| SWE-bench Verified (bash only) | 74.2% | 70.8% |
| LMArena WebDev | 1439 | 1436 |
| SWE-bench Multilingual | 68.7% | 67.3% |
| WeirdML | 69.9% | 45.6% |
| LMArena Coding | 1482 | 1474 |
| ALE-Bench | 1,177 | 821.65 |
| SciCode | — | 49% |
| GSO | 18.6% | — |
| AlgoTune | 1.83 | — |
Agentic & Tool Use Gemini 3 Pro leads
Gemini 3 Pro: 40.5 (#21), Kimi K2.5: 34.0 (#44)
| Benchmark | Gemini 3 Pro | Kimi K2.5 |
|---|---|---|
| Terminal-Bench | 69.4% | 43.2% |
| Vending-Bench 2 | 5,478 | 1,198 |
| Berkeley Function Calling Leaderboard | 72.5% | — |
| GDPval | 40.3% | — |
| Remote Labor Index | 1.3% | — |
| τ²-bench Airline | 80.5% | — |
| τ²-bench Banking | 18% | — |
| τ²-bench Retail | 75.9% | — |
| τ²-bench Telecom | 91% | — |
| DeepResearch Bench | 46.3% | — |
| OSWorld | — | 63.3% |
| BALROG | 58.1% | — |
| LMArena Search | 1207 | — |
| METR Time Horizons | 71% | — |
Reasoning Gemini 3 Pro leads
Gemini 3 Pro: 52.5 (#28), Kimi K2.5: 31.2 (#74)
| Benchmark | Gemini 3 Pro | Kimi K2.5 |
|---|---|---|
| ARC-AGI-2 | 31.1% | 11.8% |
| SimpleBench | 76.4% | 46.8% |
| Kagi LLM Benchmark | 80.1% | 78.5% |
| NYT Connections (extended) | 94.4% | 69.9% |
| ARC-AGI-1 | 75% | 65.3% |
| CritPt | 6.9% | 3.1% |
| Chess Puzzles | 31% | 12% |
| EnigmaEval | 18.2% | 3.4% |
| LMArena Hard Prompts | 1481 | 1454 |
| Epoch Capabilities Index | 152.96 | 148.05 |
| Thematic Generalization | — | 69.4% |
| ForecastBench | 61.2 | — |
Math Kimi K2.5 leads
Gemini 3 Pro: 48.8 (#50), Kimi K2.5: 50.3 (#46)
| Benchmark | Gemini 3 Pro | Kimi K2.5 |
|---|---|---|
| MathArena Final-Answer Competitions | 67% | 62.3% |
| OTIS Mock AIME 2024-2025 | 91.4% | 92.2% |
| LMArena Math | 1475 | 1470 |
| FrontierMath (Feb 2025 set) | 37.6% | 27.9% |
| FrontierMath Tier 4 (v1) | 18.8% | 4.2% |
| ProofBench | 20% | — |
| Omni-MATH | 55.5% | — |
Knowledge Gemini 3 Pro leads
Gemini 3 Pro: 64.3 (#15), Kimi K2.5: 53.6 (#52)
| Benchmark | Gemini 3 Pro | Kimi K2.5 |
|---|---|---|
| GPQA Diamond | 92.6% | 87.6% |
| Humanity's Last Exam | 37.5% | 24.4% |
| Vectara Hallucination Rate | 13.6% | 14.2% |
| LMArena Expert | 1475 | 1466 |
| SimpleQA Verified | — | 34.3% |
| MMLU-Pro | 90.3% | — |
| GPQA (HELM) | 80.3% | — |
Multimodal Gemini 3 Pro leads
Gemini 3 Pro: 56.9 (#1), Kimi K2.5: 40.4 (#37)
| Benchmark | Gemini 3 Pro | Kimi K2.5 |
|---|---|---|
| LMArena Vision | 1305 | 1269 |
| LMArena Document | 1434 | 1430 |
| GeoBench | 84% | — |
| VPCT | 91% | — |
Multilingual Gemini 3 Pro leads
Gemini 3 Pro: 57.1 (#14), Kimi K2.5: 54.1 (#50)
| Benchmark | Gemini 3 Pro | Kimi K2.5 |
|---|---|---|
| LMArena Non-English | 1474 | 1433 |
| LMArena Chinese | 1522 | 1495 |
| LMArena French | 1492 | 1454 |
| LMArena German | 1517 | 1441 |
| LMArena Japanese | 1513 | 1422 |
| LMArena Korean | 1449 | 1412 |
| LMArena Russian | 1493 | 1436 |
| LMArena Spanish | 1468 | 1450 |
Instruction Following Too close to call
Gemini 3 Pro: 76.4 (#42), Kimi K2.5: 75.5 (#60)
| Benchmark | Gemini 3 Pro | Kimi K2.5 |
|---|---|---|
| LMArena Instruction Following | 1458 | 1431 |
| IFEval | 87.7% | — |
Long Context Kimi K2.5 leads
Gemini 3 Pro: 44.0 (#75), Kimi K2.5: 52.1 (#7)
| Benchmark | Gemini 3 Pro | Kimi K2.5 |
|---|---|---|
| CL-bench | 15.8% | 19.3% |
| LMArena Longer Query | 1472 | 1445 |
| Fiction.LiveBench | — | 86.1% |
| CL-bench Life | — | 13.2% |
Writing & Preference Gemini 3 Pro leads
Gemini 3 Pro: 66.6 (#34), Kimi K2.5: 65.3 (#51)
| Benchmark | Gemini 3 Pro | Kimi K2.5 |
|---|---|---|
| LMArena Text | 1479 | 1445 |
| LMArena Creative Writing | 1482 | 1423 |
| EQ-Bench Creative Writing | 1525 | 1579 |
| LMArena Multi-Turn | 1484 | 1444 |
| WildBench | 85.9% | — |
Frequently asked questions
Is Gemini 3 Pro better than Kimi K2.5?
Gemini 3 Pro is the stronger model overall, scoring 54.7 to 47.8 on the Noometry Index.
Is Gemini 3 Pro or Kimi K2.5 better for coding?
Gemini 3 Pro scores higher on coding benchmarks: 52.1 versus 48.7 in the Noometry coding category.
How many benchmarks do Gemini 3 Pro and Kimi K2.5 share?
45 benchmarks have published results for both models. Gemini 3 Pro has 67 scored results on Noometry and Kimi K2.5 has 51.