Model comparison
gpt-oss-20b vs Kimi K2.6
Kimi K2.6 is the stronger model overall, scoring 47.7 to 32.5 on the Noometry Index. gpt-oss-20b costs 48× less per token, which makes it the better buy when Kimi K2.6's lead doesn't matter for your workload.
Last verified . 27 shared benchmarks.
Summary
- They share 27 benchmarks with published results for both. gpt-oss-20b scores higher in 0 categories and Kimi K2.6 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Kimi K2.6 leads 68.5 to 35.5.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 65.3% for gpt-oss-20b and 96.1% for Kimi K2.6.
- gpt-oss-20b is cheaper at $0.018 / $0.09 per million input/output tokens, against $0.95 / $4 for Kimi K2.6.
- Kimi K2.6 accepts more context: 262K tokens versus 131K.
Side by side
| gpt-oss-20b | Kimi K2.6 | |
|---|---|---|
| Provider | OpenAI | Moonshot AI |
| Noometry Index | 32.5 | 47.7 |
| Released | 2025-08-05 | 2026-04-20 |
| Weights | Open | Open |
| Context window | 131K | 262K |
| Max output | 16K | 262K |
| Input $ / M tokens | $0.018 | $0.95 |
| Output $ / M tokens | $0.09 | $4 |
| Results tracked | 34 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K2.6 leads
gpt-oss-20b: 37.6 (#192), Kimi K2.6: 50.7 (#43)
| Benchmark | gpt-oss-20b | Kimi K2.6 |
|---|---|---|
| SciCode | 34.4% | 53.5% |
| WeirdML | 40.9% | 55.9% |
| LMArena Coding | 1306 | 1488 |
| ALE-Bench | 566.05 | 1,093 |
| SWE-bench Verified | — | 76.7% |
| LMArena WebDev | — | 1509 |
Agentic & Tool Use Kimi K2.6 leads
gpt-oss-20b: 9.3 (#154), Kimi K2.6: 21.9 (#137)
| Benchmark | gpt-oss-20b | Kimi K2.6 |
|---|---|---|
| Terminal-Bench | 3.4% | — |
| OSWorld 2.0 | — | 4.6% |
| ExploitBench | — | 18.4% |
| GBAEval | — | 0.9% |
| GDP.pdf | — | 12% |
| Vending-Bench 2 | — | 6,205 |
Reasoning Kimi K2.6 leads
gpt-oss-20b: 19.3 (#261), Kimi K2.6: 40.5 (#55)
| Benchmark | gpt-oss-20b | Kimi K2.6 |
|---|---|---|
| CritPt | 1.4% | 8% |
| Chess Puzzles | 4% | 26% |
| LMArena Hard Prompts | 1274 | 1470 |
| DTBench | 68% | 90.9% |
| LMCA | 14.5% | 37.3% |
| Epoch Capabilities Index | 137.82 | 151.05 |
| Kagi LLM Benchmark | 53.2% | — |
| NYT Connections (extended) | — | 87.2% |
| EBR-Bench | — | 2.4% |
| Mystery Game Puzzles | — | 18% |
Math Kimi K2.6 leads
gpt-oss-20b: 39.4 (#103), Kimi K2.6: 57.0 (#41)
| Benchmark | gpt-oss-20b | Kimi K2.6 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 65.3% | 96.1% |
| LMArena Math | 1317 | 1475 |
| FrontierMath (Tiers 1-3) | — | 57.2% |
| FrontierMath Tier 4 | — | 25.6% |
| MathArena Final-Answer Competitions | — | 72.9% |
| ProofBench | — | 16% |
| Omni-MATH | 56.5% | — |
| FrontierMath (Feb 2025 set) | — | 39% |
| FrontierMath Tier 4 (v1) | — | 14.6% |
Knowledge Kimi K2.6 leads
gpt-oss-20b: 34.6 (#195), Kimi K2.6: 54.0 (#54)
| Benchmark | gpt-oss-20b | Kimi K2.6 |
|---|---|---|
| GPQA Diamond | 60.8% | 90.8% |
| LMArena Expert | 1258 | 1491 |
| SimpleQA Verified | — | 34.9% |
| MMLU-Pro | 74% | — |
| Vectara Hallucination Rate | — | 10.8% |
| GPQA (HELM) | 59.4% | — |
Multimodal Not comparable
gpt-oss-20b: —, Kimi K2.6: 31.6 (#103)
| Benchmark | gpt-oss-20b | Kimi K2.6 |
|---|---|---|
| LMArena Vision | — | 1283 |
| Blueprint-Bench 2 | — | 3.9% |
| Furniture Assembly | — | 21.7% |
| LMArena Document | — | 1451 |
Multilingual Kimi K2.6 leads
gpt-oss-20b: 42.2 (#197), Kimi K2.6: 54.9 (#37)
| Benchmark | gpt-oss-20b | Kimi K2.6 |
|---|---|---|
| LMArena Non-English | 1268 | 1446 |
| LMArena Chinese | 1314 | 1521 |
| LMArena German | 1255 | 1450 |
| LMArena Japanese | 1244 | 1443 |
| LMArena Korean | 1236 | 1427 |
| LMArena Russian | 1278 | 1446 |
| LMArena Spanish | 1267 | 1464 |
| LMArena French | — | 1471 |
Instruction Following Kimi K2.6 leads
gpt-oss-20b: 61.8 (#240), Kimi K2.6: 76.3 (#43)
| Benchmark | gpt-oss-20b | Kimi K2.6 |
|---|---|---|
| LMArena Instruction Following | 1236 | 1451 |
| IFEval | 73.2% | — |
Long Context Kimi K2.6 leads
gpt-oss-20b: 37.9 (#209), Kimi K2.6: 44.9 (#52)
| Benchmark | gpt-oss-20b | Kimi K2.6 |
|---|---|---|
| LMArena Longer Query | 1250 | 1468 |
Writing & Preference Kimi K2.6 leads
gpt-oss-20b: 35.5 (#265), Kimi K2.6: 68.5 (#26)
| Benchmark | gpt-oss-20b | Kimi K2.6 |
|---|---|---|
| LMArena Text | 1287 | 1455 |
| LMArena Creative Writing | 1201 | 1434 |
| EQ-Bench Creative Writing | 666 | 1725 |
| LMArena Multi-Turn | 1268 | 1453 |
| WildBench | 73.7% | — |
| EQ-Bench 4 | — | 1202 |
Frequently asked questions
Is gpt-oss-20b better than Kimi K2.6?
Kimi K2.6 is the stronger model overall, scoring 47.7 to 32.5 on the Noometry Index. gpt-oss-20b costs 48× less per token, which makes it the better buy when Kimi K2.6's lead doesn't matter for your workload.
Which is cheaper, gpt-oss-20b or Kimi K2.6?
gpt-oss-20b is cheaper. It lists at $0.018 per million input tokens and $0.09 per million output tokens; Kimi K2.6 lists at $0.95 and $4.
Is gpt-oss-20b or Kimi K2.6 better for coding?
Kimi K2.6 scores higher on coding benchmarks: 50.7 versus 37.6 in the Noometry coding category.
Which has the bigger context window?
Kimi K2.6 does, with 262K tokens against 131K.
How many benchmarks do gpt-oss-20b and Kimi K2.6 share?
27 benchmarks have published results for both models. gpt-oss-20b has 34 scored results on Noometry and Kimi K2.6 has 51.