# GPT-4.1 vs Kimi K2.5

> Kimi K2.5 is the stronger model overall, scoring 47.8 to 35.7 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gpt-4-1-vs-kimi-k2-5
- Last updated: 2026-10-10
- Shared benchmarks: 38

## Summary

- They share 38 benchmarks with published results for both. GPT-4.1 scores higher in 1 category and Kimi K2.5 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K2.5 leads 50.3 to 21.4.
- The biggest single-benchmark swing is ARC-AGI-1: 5.5% for GPT-4.1 and 65.3% for Kimi K2.5.
- Kimi K2.5 is cheaper at $0.45 / $2.25 per million input/output tokens, against $2 / $8 for GPT-4.1.
- GPT-4.1 accepts more context: 1.05M tokens versus 262K.
- Kimi K2.5 has downloadable open weights; the other is API-only.

## Snapshot

| | GPT-4.1 | Kimi K2.5 |
|---|---|---|
| Provider | OpenAI | Moonshot AI |
| Noometry Index | 35.7 | 47.8 |
| Rank | 215 | 50 |
| Context | 1.05M | 262K |
| Input $/M | $2 | $0.45 |
| Output $/M | $8 | $2.25 |
| Weights | Proprietary | Open |

## Coding

- GPT-4.1: 34.2 (#228)
- Kimi K2.5: 48.7 (#48)

| Benchmark | GPT-4.1 | Kimi K2.5 |
|---|---|---|
| SWE-bench Verified | 48.5% | 73.8% |
| SWE-bench Verified (bash only) | 39.6% | 70.8% |
| WeirdML | 39% | 45.6% |
| LMArena Coding | 1391 | 1474 |
| ALE-Bench | 558.1 | 821.65 |
| Aider Polyglot | 52.4% | — |
| LMArena WebDev | — | 1436 |
| SWE-bench Multilingual | — | 67.3% |
| SciCode | — | 49% |
| CadEval | 42% | — |

## Agentic & Tool Use

- GPT-4.1: 34.4 (#40)
- Kimi K2.5: 34.0 (#44)

| Benchmark | GPT-4.1 | Kimi K2.5 |
|---|---|---|
| Terminal-Bench | — | 43.2% |
| Berkeley Function Calling Leaderboard | 54% | — |
| OSWorld | — | 63.3% |
| Vending-Bench 2 | — | 1,198 |

## Reasoning

- GPT-4.1: 11.7 (#334)
- Kimi K2.5: 31.2 (#74)

| Benchmark | GPT-4.1 | Kimi K2.5 |
|---|---|---|
| ARC-AGI-2 | 0.4% | 11.8% |
| SimpleBench | 27% | 46.8% |
| Kagi LLM Benchmark | 52.3% | 78.5% |
| ARC-AGI-1 | 5.5% | 65.3% |
| Chess Puzzles | 6% | 12% |
| EnigmaEval | 2.2% | 3.4% |
| LMArena Hard Prompts | 1384 | 1454 |
| Epoch Capabilities Index | 136.79 | 148.05 |
| NYT Connections (extended) | — | 69.9% |
| CritPt | — | 3.1% |
| Thematic Generalization | — | 69.4% |
| DTBench | 68.3% | — |
| LMCA | 25.6% | — |
| ForecastBench | 61.5 | — |

## Math

- GPT-4.1: 21.4 (#269)
- Kimi K2.5: 50.3 (#46)

| Benchmark | GPT-4.1 | Kimi K2.5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 38.3% | 92.2% |
| LMArena Math | 1369 | 1470 |
| FrontierMath (Feb 2025 set) | 5.5% | 27.9% |
| FrontierMath Tier 4 (v1) | 0% | 4.2% |
| FrontierMath (Tiers 1-3) | 6% | — |
| MathArena Final-Answer Competitions | — | 62.3% |
| Omni-MATH | 47.1% | — |
| MATH Level 5 | 83% | — |

## Knowledge

- GPT-4.1: 37.0 (#152)
- Kimi K2.5: 53.6 (#52)

| Benchmark | GPT-4.1 | Kimi K2.5 |
|---|---|---|
| GPQA Diamond | 66.9% | 87.6% |
| Humanity's Last Exam | 5.4% | 24.4% |
| SimpleQA Verified | 31.1% | 34.3% |
| Vectara Hallucination Rate | 5.6% | 14.2% |
| LMArena Expert | 1364 | 1466 |
| MMLU-Pro | 81.1% | — |
| GPQA (HELM) | 65.9% | — |

## Multimodal

- GPT-4.1: 37.5 (#66)
- Kimi K2.5: 40.4 (#37)

| Benchmark | GPT-4.1 | Kimi K2.5 |
|---|---|---|
| LMArena Vision | 1210 | 1269 |
| GeoBench | 72% | — |
| LMArena Document | — | 1430 |

## Multilingual

- GPT-4.1: 49.6 (#128)
- Kimi K2.5: 54.1 (#50)

| Benchmark | GPT-4.1 | Kimi K2.5 |
|---|---|---|
| LMArena Non-English | 1370 | 1433 |
| LMArena Chinese | 1382 | 1495 |
| LMArena French | 1383 | 1454 |
| LMArena German | 1380 | 1441 |
| LMArena Japanese | 1320 | 1422 |
| LMArena Korean | 1339 | 1412 |
| LMArena Russian | 1377 | 1436 |
| LMArena Spanish | 1375 | 1450 |

## Instruction Following

- GPT-4.1: 71.5 (#148)
- Kimi K2.5: 75.5 (#60)

| Benchmark | GPT-4.1 | Kimi K2.5 |
|---|---|---|
| LMArena Instruction Following | 1367 | 1431 |
| IFEval | 83.8% | — |

## Long Context

- GPT-4.1: 40.0 (#158)
- Kimi K2.5: 52.1 (#7)

| Benchmark | GPT-4.1 | Kimi K2.5 |
|---|---|---|
| Fiction.LiveBench | 63.9% | 86.1% |
| LMArena Longer Query | 1385 | 1445 |
| CL-bench | — | 19.3% |
| CL-bench Life | — | 13.2% |

## Writing & Preference

- GPT-4.1: 57.8 (#120)
- Kimi K2.5: 65.3 (#51)

| Benchmark | GPT-4.1 | Kimi K2.5 |
|---|---|---|
| LMArena Text | 1383 | 1445 |
| LMArena Creative Writing | 1363 | 1423 |
| EQ-Bench Creative Writing | 1420 | 1579 |
| LMArena Multi-Turn | 1398 | 1444 |
| WildBench | 85.4% | — |

## FAQ

### Is GPT-4.1 better than Kimi K2.5?

Kimi K2.5 is the stronger model overall, scoring 47.8 to 35.7 on the Noometry Index.

### Which is cheaper, GPT-4.1 or Kimi K2.5?

Kimi K2.5 is cheaper. It lists at $0.45 per million input tokens and $2.25 per million output tokens; GPT-4.1 lists at $2 and $8.

### Is GPT-4.1 or Kimi K2.5 better for coding?

Kimi K2.5 scores higher on coding benchmarks: 48.7 versus 34.2 in the Noometry coding category.

### Which has the bigger context window?

GPT-4.1 does, with 1.05M tokens against 262K.

### How many benchmarks do GPT-4.1 and Kimi K2.5 share?

38 benchmarks have published results for both models. GPT-4.1 has 52 scored results on Noometry and Kimi K2.5 has 51.