# gpt-oss-20b vs Kimi K2 (Jul 2025)

> Kimi K2 (Jul 2025) is the stronger model overall, scoring 40.9 to 32.1 on the Noometry Index. gpt-oss-20b costs 28× less per token, which makes it the better buy when Kimi K2 (Jul 2025)'s lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/gpt-oss-20b-vs-kimi-k2
- Last updated: 2026-10-10
- Shared benchmarks: 27

## Summary

- They share 27 benchmarks with published results for both. gpt-oss-20b scores higher in 0 categories and Kimi K2 (Jul 2025) in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Kimi K2 (Jul 2025) leads 62.5 to 35.8.
- The biggest single-benchmark swing is Terminal-Bench: 3.4% for gpt-oss-20b and 35.7% for Kimi K2 (Jul 2025).
- gpt-oss-20b is cheaper at $0.018 / $0.09 per million input/output tokens, against $0.57 / $2.30 for Kimi K2 (Jul 2025).
- Kimi K2 (Jul 2025) accepts more context: 262K tokens versus 131K.

## Snapshot

| | gpt-oss-20b | Kimi K2 (Jul 2025) |
|---|---|---|
| Provider | OpenAI | Moonshot AI |
| Noometry Index | 32.1 | 40.9 |
| Rank | 249 | 134 |
| Context | 131K | 262K |
| Input $/M | $0.018 | $0.57 |
| Output $/M | $0.09 | $2.30 |
| Weights | Open | Open |

## Coding

- gpt-oss-20b: 37.1 (#185)
- Kimi K2 (Jul 2025): 43.1 (#83)

| Benchmark | gpt-oss-20b | Kimi K2 (Jul 2025) |
|---|---|---|
| WeirdML | 40.9% | 42.8% |
| LMArena Coding | 1307 | 1399 |
| ALE-Bench | 566.05 | 597.5 |
| SWE-bench Verified (bash only) | — | 63.4% |
| Aider Polyglot | — | 59.1% |
| SciCode | 34.4% | — |
| GSO | — | 4.9% |

## Agentic & Tool Use

- gpt-oss-20b: 9.2 (#146)
- Kimi K2 (Jul 2025): 32.2 (#59)

| Benchmark | gpt-oss-20b | Kimi K2 (Jul 2025) |
|---|---|---|
| Terminal-Bench | 3.4% | 35.7% |
| Berkeley Function Calling Leaderboard | — | 59.1% |
| METR Time Horizons | — | 59.2% |

## Reasoning

- gpt-oss-20b: 18.8 (#251)
- Kimi K2 (Jul 2025): 22.7 (#173)

| Benchmark | gpt-oss-20b | Kimi K2 (Jul 2025) |
|---|---|---|
| Kagi LLM Benchmark | 53.2% | 64.4% |
| LMArena Hard Prompts | 1274 | 1384 |
| Epoch Capabilities Index | 137.81 | 146.02 |
| SimpleBench | — | 26.3% |
| CritPt | 1.4% | — |
| Chess Puzzles | 4% | — |
| DTBench | 68% | — |
| LMCA | 14.5% | — |
| ForecastBench | — | 60.2 |

## Math

- gpt-oss-20b: 37.6 (#93)
- Kimi K2 (Jul 2025): 40.1 (#81)

| Benchmark | gpt-oss-20b | Kimi K2 (Jul 2025) |
|---|---|---|
| Omni-MATH | 56.5% | 65.4% |
| LMArena Math | 1317 | 1396 |
| OTIS Mock AIME 2024-2025 | 65.3% | — |
| FrontierMath (Feb 2025 set) | — | 21.4% |
| FrontierMath Tier 4 (v1) | — | 0% |

## Knowledge

- gpt-oss-20b: 34.6 (#186)
- Kimi K2 (Jul 2025): 36.9 (#153)

| Benchmark | gpt-oss-20b | Kimi K2 (Jul 2025) |
|---|---|---|
| MMLU-Pro | 74% | 81.9% |
| GPQA (HELM) | 59.4% | 65.3% |
| LMArena Expert | 1258 | 1363 |
| GPQA Diamond | 60.8% | — |
| Confabulations | — | 20.4% |
| Vectara Hallucination Rate | — | 17.9% |

## Multilingual

- gpt-oss-20b: 42.4 (#192)
- Kimi K2 (Jul 2025): 49.8 (#125)

| Benchmark | gpt-oss-20b | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Non-English | 1269 | 1372 |
| LMArena Chinese | 1311 | 1416 |
| LMArena German | 1249 | 1386 |
| LMArena Japanese | 1245 | 1349 |
| LMArena Korean | 1239 | 1324 |
| LMArena Russian | 1279 | 1384 |
| LMArena Spanish | 1267 | 1383 |
| LMArena French | — | 1379 |

## Instruction Following

- gpt-oss-20b: 62.0 (#235)
- Kimi K2 (Jul 2025): 71.3 (#151)

| Benchmark | gpt-oss-20b | Kimi K2 (Jul 2025) |
|---|---|---|
| IFEval | 73.2% | 85% |
| LMArena Instruction Following | 1236 | 1347 |

## Long Context

- gpt-oss-20b: 37.9 (#203)
- Kimi K2 (Jul 2025): 41.2 (#140)

| Benchmark | gpt-oss-20b | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Longer Query | 1251 | 1353 |
| Fiction.LiveBench | — | 66.7% |
| CL-bench | — | 17.6% |

## Writing & Preference

- gpt-oss-20b: 35.8 (#260)
- Kimi K2 (Jul 2025): 62.5 (#76)

| Benchmark | gpt-oss-20b | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Text | 1288 | 1380 |
| LMArena Creative Writing | 1201 | 1349 |
| EQ-Bench Creative Writing | 666 | 1666 |
| WildBench | 73.7% | 86.2% |
| LMArena Multi-Turn | 1268 | 1371 |
| Short-Story Creative Writing | — | 85.6% |

## FAQ

### Is gpt-oss-20b better than Kimi K2 (Jul 2025)?

Kimi K2 (Jul 2025) is the stronger model overall, scoring 40.9 to 32.1 on the Noometry Index. gpt-oss-20b costs 28× less per token, which makes it the better buy when Kimi K2 (Jul 2025)'s lead doesn't matter for your workload.

### Which is cheaper, gpt-oss-20b or Kimi K2 (Jul 2025)?

gpt-oss-20b is cheaper. It lists at $0.018 per million input tokens and $0.09 per million output tokens; Kimi K2 (Jul 2025) lists at $0.57 and $2.30.

### Is gpt-oss-20b or Kimi K2 (Jul 2025) better for coding?

Kimi K2 (Jul 2025) scores higher on coding benchmarks: 43.1 versus 37.1 in the Noometry coding category.

### Which has the bigger context window?

Kimi K2 (Jul 2025) does, with 262K tokens against 131K.

### How many benchmarks do gpt-oss-20b and Kimi K2 (Jul 2025) share?

27 benchmarks have published results for both models. gpt-oss-20b has 34 scored results on Noometry and Kimi K2 (Jul 2025) has 42.