# Gemini 2.5 Pro vs Kimi K3

> Kimi K3 is the stronger model overall, scoring 59.5 to 45.0 on the Noometry Index. Gemini 2.5 Pro costs 1.7× less per token, which makes it the better buy when Kimi K3's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/gemini-2-5-pro-vs-kimi-k3
- Last updated: 2026-10-10
- Shared benchmarks: 37

## Summary

- They share 37 benchmarks with published results for both. Gemini 2.5 Pro scores higher in 2 categories and Kimi K3 in 8 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K3 leads 74.2 to 32.5.
- The biggest single-benchmark swing is ARC-AGI-2: 4.9% for Gemini 2.5 Pro and 60.4% for Kimi K3.
- Gemini 2.5 Pro is cheaper at $1.25 / $10 per million input/output tokens, against $3 / $15 for Kimi K3.
- Kimi K3 has downloadable open weights; the other is API-only.

## Snapshot

| | Gemini 2.5 Pro | Kimi K3 |
|---|---|---|
| Provider | Google | Moonshot AI |
| Noometry Index | 45.0 | 59.5 |
| Rank | 75 | 15 |
| Context | 1.05M | 1.05M |
| Input $/M | $1.25 | $3 |
| Output $/M | $10 | $15 |
| Weights | Proprietary | Open |

## Coding

- Gemini 2.5 Pro: 42.4 (#101)
- Kimi K3: 61.0 (#10)

| Benchmark | Gemini 2.5 Pro | Kimi K3 |
|---|---|---|
| LMArena WebDev | 1227 | 1654 |
| SciCode | 42.8% | 59.5% |
| WeirdML | 54% | 82.6% |
| LMArena Coding | 1452 | 1508 |
| ALE-Bench | 785.52 | 1,524 |
| SWE-bench Verified | 57.6% | — |
| DeepSWE | — | 68.5% |
| FrontierCode | — | 44.2% |
| SWE-bench Verified (bash only) | 53.6% | — |
| Aider Polyglot | 83.1% | — |
| FrontierSWE | — | 25.9% |
| GSO | 3.9% | — |
| LiveBench Coding | 85.9% | — |
| CadEval | 64% | — |
| AlgoTune | 1.51 | — |

## Agentic & Tool Use

- Gemini 2.5 Pro: 29.2 (#88)
- Kimi K3: 41.8 (#20)

| Benchmark | Gemini 2.5 Pro | Kimi K3 |
|---|---|---|
| τ²-bench Banking | 13.7% | 37.1% |
| Vending-Bench 2 | 573.64 | 5,165 |
| Terminal-Bench | 32.6% | — |
| APEX-Agents | — | 50.6% |
| GDPval | 23.3% | — |
| Remote Labor Index | 0.8% | — |
| TheAgentCompany | 30.3% | — |
| DeepResearch Bench | 42.8% | — |
| PostTrainBench | — | 32% |
| BALROG | 43.3% | — |
| GBAEval | — | 48.3% |
| GDP.pdf | — | 19% |
| LMArena Search | 1142 | — |
| METR Time Horizons | 55.4% | — |

## Reasoning

- Gemini 2.5 Pro: 28.8 (#99)
- Kimi K3: 63.0 (#17)

| Benchmark | Gemini 2.5 Pro | Kimi K3 |
|---|---|---|
| ARC-AGI-2 | 4.9% | 60.4% |
| SimpleBench | 62.4% | 60.7% |
| ARC-AGI-1 | 41% | 94.5% |
| CritPt | 2% | 23.4% |
| Chess Puzzles | 20% | 39% |
| LMArena Hard Prompts | 1455 | 1496 |
| DTBench | 82.4% | 91.2% |
| LMCA | 34.8% | 52.7% |
| Epoch Capabilities Index | 145.32 | 157.45 |
| ForecastBench | 61.3 | 61.1 |
| Kagi LLM Benchmark | 70.3% | — |
| NYT Connections (extended) | — | 93.6% |
| EnigmaEval | 5.6% | — |
| LiveBench Reasoning | 89.8% | — |
| Mystery Game Puzzles | — | 26% |
| LiveBench Data Analysis | 79.9% | — |
| Surface Evolver Bench | — | 95% |
| LiveBench | 82.3% | — |

## Math

- Gemini 2.5 Pro: 32.5 (#213)
- Kimi K3: 74.2 (#16)

| Benchmark | Gemini 2.5 Pro | Kimi K3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 24.6% | 72.2% |
| FrontierMath Tier 4 | 0% | 39% |
| OTIS Mock AIME 2024-2025 | 84.7% | 97.2% |
| LMArena Math | 1450 | 1491 |
| MathArena Final-Answer Competitions | — | 87.8% |
| ProofBench | — | 87% |
| Omni-MATH | 41.6% | — |
| LiveBench Math | 90.2% | — |
| MATH Level 5 | 95.9% | — |
| FrontierMath (Feb 2025 set) | 14.1% | — |
| FrontierMath Tier 4 (v1) | 4.2% | — |

## Knowledge

- Gemini 2.5 Pro: 56.0 (#46)
- Kimi K3: 63.2 (#21)

| Benchmark | Gemini 2.5 Pro | Kimi K3 |
|---|---|---|
| GPQA Diamond | 85.3% | 93.1% |
| LMArena Expert | 1452 | 1521 |
| Humanity's Last Exam | 21.6% | — |
| SimpleQA Verified | — | 50.6% |
| MMLU-Pro | 86.3% | — |
| Confabulations | 10.6% | — |
| Vectara Hallucination Rate | 7% | — |
| GPQA (HELM) | 74.9% | — |

## Multimodal

- Gemini 2.5 Pro: 45.2 (#18)
- Kimi K3: 37.8 (#70)

| Benchmark | Gemini 2.5 Pro | Kimi K3 |
|---|---|---|
| LMArena Vision | 1263 | — |
| GeoBench | 86% | — |
| VPCT | 48% | — |
| Blueprint-Bench 2 | — | 29.5% |
| Furniture Assembly | — | 34.2% |
| LMArena Document | 1421 | — |
| SpatialViz-Bench | 44.7% | — |

## Multilingual

- Gemini 2.5 Pro: 55.3 (#31)
- Kimi K3: 56.3 (#21)

| Benchmark | Gemini 2.5 Pro | Kimi K3 |
|---|---|---|
| LMArena Non-English | 1451 | 1466 |
| LMArena Chinese | 1507 | 1529 |
| LMArena French | 1472 | 1491 |
| LMArena German | 1487 | 1488 |
| LMArena Japanese | 1461 | 1487 |
| LMArena Korean | 1434 | 1458 |
| LMArena Russian | 1461 | 1482 |
| LMArena Spanish | 1473 | 1472 |

## Instruction Following

- Gemini 2.5 Pro: 75.0 (#75)
- Kimi K3: 77.7 (#14)

| Benchmark | Gemini 2.5 Pro | Kimi K3 |
|---|---|---|
| LMArena Instruction Following | 1437 | 1483 |
| LiveBench Instruction Following | 80.6% | — |
| IFEval | 84% | — |

## Long Context

- Gemini 2.5 Pro: 59.8 (#5)
- Kimi K3: 45.8 (#29)

| Benchmark | Gemini 2.5 Pro | Kimi K3 |
|---|---|---|
| LMArena Longer Query | 1449 | 1494 |
| Fiction.LiveBench | 91.7% | — |

## Writing & Preference

- Gemini 2.5 Pro: 63.7 (#62)
- Kimi K3: 76.6 (#4)

| Benchmark | Gemini 2.5 Pro | Kimi K3 |
|---|---|---|
| LMArena Text | 1458 | 1476 |
| LMArena Creative Writing | 1454 | 1454 |
| EQ-Bench Creative Writing | 1421 | 2082 |
| LMArena Multi-Turn | 1453 | 1488 |
| Short-Story Creative Writing | 83.8% | — |
| WildBench | 85.7% | — |
| EQ-Bench 4 | — | 1339 |
| LiveBench Language | 67.8% | — |

## FAQ

### Is Gemini 2.5 Pro better than Kimi K3?

Kimi K3 is the stronger model overall, scoring 59.5 to 45.0 on the Noometry Index. Gemini 2.5 Pro costs 1.7× less per token, which makes it the better buy when Kimi K3's lead doesn't matter for your workload.

### Which is cheaper, Gemini 2.5 Pro or Kimi K3?

Gemini 2.5 Pro is cheaper. It lists at $1.25 per million input tokens and $10 per million output tokens; Kimi K3 lists at $3 and $15.

### Is Gemini 2.5 Pro or Kimi K3 better for coding?

Kimi K3 scores higher on coding benchmarks: 61.0 versus 42.4 in the Noometry coding category.

### Which has the bigger context window?

Both accept 1.05M tokens.

### How many benchmarks do Gemini 2.5 Pro and Kimi K3 share?

37 benchmarks have published results for both models. Gemini 2.5 Pro has 78 scored results on Noometry and Kimi K3 has 53.