# GPT-5 Nano vs Kimi K3

> Kimi K3 is the stronger model overall, scoring 59.5 to 33.4 on the Noometry Index. GPT-5 Nano costs 44× 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/gpt-5-nano-vs-kimi-k3
- Last updated: 2026-10-10
- Shared benchmarks: 32

## Summary

- They share 32 benchmarks with published results for both. GPT-5 Nano scores higher in 0 categories and Kimi K3 in 10 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Kimi K3 leads 63.0 to 16.4.
- The biggest single-benchmark swing is ProofBench: 12% for GPT-5 Nano and 87% for Kimi K3.
- GPT-5 Nano is cheaper at $0.05 / $0.40 per million input/output tokens, against $3 / $15 for Kimi K3.
- Kimi K3 accepts more context: 1.05M tokens versus 400K.
- Kimi K3 has downloadable open weights; the other is API-only.

## Snapshot

| | GPT-5 Nano | Kimi K3 |
|---|---|---|
| Provider | OpenAI | Moonshot AI |
| Noometry Index | 33.4 | 59.5 |
| Rank | 234 | 12 |
| Context | 400K | 1.05M |
| Input $/M | $0.05 | $3 |
| Output $/M | $0.40 | $15 |
| Weights | Proprietary | Open |

## Coding

- GPT-5 Nano: 33.3 (#248)
- Kimi K3: 60.6 (#8)

| Benchmark | GPT-5 Nano | Kimi K3 |
|---|---|---|
| WeirdML | 38.1% | 82.6% |
| LMArena Coding | 1351 | 1509 |
| ALE-Bench | 718.67 | 1,524 |
| DeepSWE | — | 68.5% |
| FrontierCode | — | 44.2% |
| SWE-bench Verified (bash only) | 34.8% | — |
| LMArena WebDev | — | 1659 |
| FrontierSWE | — | 25.9% |
| SciCode | — | 59.5% |

## Agentic & Tool Use

- GPT-5 Nano: 25.6 (#99)
- Kimi K3: 41.8 (#19)

| Benchmark | GPT-5 Nano | Kimi K3 |
|---|---|---|
| Terminal-Bench | 21.8% | — |
| APEX-Agents | — | 50.6% |
| Berkeley Function Calling Leaderboard | 51.5% | — |
| τ²-bench Banking | — | 37.1% |
| PostTrainBench | — | 32% |
| GBAEval | — | 48.3% |
| GDP.pdf | — | 19% |
| Vending-Bench 2 | — | 5,165 |

## Reasoning

- GPT-5 Nano: 16.4 (#291)
- Kimi K3: 63.0 (#15)

| Benchmark | GPT-5 Nano | Kimi K3 |
|---|---|---|
| ARC-AGI-2 | 2.6% | 60.4% |
| ARC-AGI-1 | 20.7% | 94.5% |
| Chess Puzzles | 27% | 39% |
| LMArena Hard Prompts | 1329 | 1495 |
| Mystery Game Puzzles | 9% | 26% |
| DTBench | 62.7% | 91.2% |
| LMCA | 7.9% | 52.7% |
| Epoch Capabilities Index | 139.39 | 157.68 |
| ForecastBench | 59.1 | 61.1 |
| SimpleBench | — | 60.7% |
| Kagi LLM Benchmark | 62.2% | — |
| NYT Connections (extended) | — | 93.6% |
| CritPt | — | 23.4% |
| Surface Evolver Bench | — | 95% |

## Math

- GPT-5 Nano: 29.1 (#221)
- Kimi K3: 74.0 (#13)

| Benchmark | GPT-5 Nano | Kimi K3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 20% | 72.2% |
| FrontierMath Tier 4 | 2.4% | 39% |
| OTIS Mock AIME 2024-2025 | 81.1% | 97.2% |
| ProofBench | 12% | 87% |
| LMArena Math | 1318 | 1493 |
| MathArena Final-Answer Competitions | — | 87.8% |
| Omni-MATH | 54.6% | — |
| MATH Level 5 | 95.2% | — |
| FrontierMath (Feb 2025 set) | 8.3% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |

## Knowledge

- GPT-5 Nano: 35.8 (#166)
- Kimi K3: 63.2 (#19)

| Benchmark | GPT-5 Nano | Kimi K3 |
|---|---|---|
| GPQA Diamond | 69.4% | 93.1% |
| SimpleQA Verified | 11.7% | 50.6% |
| LMArena Expert | 1321 | 1520 |
| MMLU-Pro | 77.8% | — |
| Vectara Hallucination Rate | 10.5% | — |
| GPQA (HELM) | 67.9% | — |

## Multimodal

- GPT-5 Nano: 30.7 (#102)
- Kimi K3: 38.4 (#58)

| Benchmark | GPT-5 Nano | Kimi K3 |
|---|---|---|
| LMArena Vision | 1159 | — |
| VPCT | 37.2% | — |
| Blueprint-Bench 2 | — | 29.5% |
| Furniture Assembly | — | 34.2% |

## Multilingual

- GPT-5 Nano: 45.5 (#167)
- Kimi K3: 56.5 (#20)

| Benchmark | GPT-5 Nano | Kimi K3 |
|---|---|---|
| LMArena Non-English | 1313 | 1466 |
| LMArena Chinese | 1355 | 1528 |
| LMArena German | 1329 | 1474 |
| LMArena Korean | 1269 | 1465 |
| LMArena Russian | 1296 | 1481 |
| LMArena Spanish | 1360 | 1478 |
| LMArena French | — | 1488 |
| LMArena Japanese | — | 1492 |

## Instruction Following

- GPT-5 Nano: 75.2 (#72)
- Kimi K3: 77.8 (#13)

| Benchmark | GPT-5 Nano | Kimi K3 |
|---|---|---|
| LMArena Instruction Following | 1306 | 1481 |
| IFEval | 93.2% | — |

## Long Context

- GPT-5 Nano: 31.3 (#275)
- Kimi K3: 45.8 (#27)

| Benchmark | GPT-5 Nano | Kimi K3 |
|---|---|---|
| LMArena Longer Query | 1313 | 1494 |
| Fiction.LiveBench | 44.4% | — |

## Writing & Preference

- GPT-5 Nano: 39.4 (#244)
- Kimi K3: 76.7 (#4)

| Benchmark | GPT-5 Nano | Kimi K3 |
|---|---|---|
| LMArena Text | 1320 | 1476 |
| LMArena Creative Writing | 1250 | 1452 |
| EQ-Bench Creative Writing | 705 | 2082 |
| LMArena Multi-Turn | 1311 | 1488 |
| WildBench | 80.6% | — |
| EQ-Bench 4 | — | 1339 |

## FAQ

### Is GPT-5 Nano better than Kimi K3?

Kimi K3 is the stronger model overall, scoring 59.5 to 33.4 on the Noometry Index. GPT-5 Nano costs 44× less per token, which makes it the better buy when Kimi K3's lead doesn't matter for your workload.

### Which is cheaper, GPT-5 Nano or Kimi K3?

GPT-5 Nano is cheaper. It lists at $0.05 per million input tokens and $0.40 per million output tokens; Kimi K3 lists at $3 and $15.

### Is GPT-5 Nano or Kimi K3 better for coding?

Kimi K3 scores higher on coding benchmarks: 60.6 versus 33.3 in the Noometry coding category.

### Which has the bigger context window?

Kimi K3 does, with 1.05M tokens against 400K.

### How many benchmarks do GPT-5 Nano and Kimi K3 share?

32 benchmarks have published results for both models. GPT-5 Nano has 48 scored results on Noometry and Kimi K3 has 53.