Model comparison
gpt-oss-20b vs Kimi K2.5
Kimi K2.5 is the stronger model overall, scoring 48.1 to 32.5 on the Noometry Index. gpt-oss-20b costs 25× less per token, which makes it the better buy when Kimi K2.5'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 K2.5 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Kimi K2.5 leads 65.1 to 35.5.
- The biggest single-benchmark swing is Terminal-Bench: 3.4% for gpt-oss-20b and 43.2% for Kimi K2.5.
- gpt-oss-20b is cheaper at $0.018 / $0.09 per million input/output tokens, against $0.45 / $2.25 for Kimi K2.5.
- Kimi K2.5 accepts more context: 262K tokens versus 131K.
Side by side
| gpt-oss-20b | Kimi K2.5 | |
|---|---|---|
| Provider | OpenAI | Moonshot AI |
| Noometry Index | 32.5 | 48.1 |
| Released | 2025-08-05 | 2026-01-27 |
| Weights | Open | Open |
| Context window | 131K | 262K |
| Max output | 16K | 262K |
| Input $ / M tokens | $0.018 | $0.45 |
| Output $ / M tokens | $0.09 | $2.25 |
| Results tracked | 34 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K2.5 leads
gpt-oss-20b: 37.6 (#192), Kimi K2.5: 48.8 (#53)
| Benchmark | gpt-oss-20b | Kimi K2.5 |
|---|---|---|
| SciCode | 34.4% | 49% |
| WeirdML | 40.9% | 45.6% |
| LMArena Coding | 1306 | 1474 |
| ALE-Bench | 566.05 | 821.65 |
| SWE-bench Verified | — | 73.8% |
| SWE-bench Verified (bash only) | — | 70.8% |
| LMArena WebDev | — | 1437 |
| SWE-bench Multilingual | — | 67.3% |
Agentic & Tool Use Kimi K2.5 leads
gpt-oss-20b: 9.3 (#154), Kimi K2.5: 34.2 (#48)
| Benchmark | gpt-oss-20b | Kimi K2.5 |
|---|---|---|
| Terminal-Bench | 3.4% | 43.2% |
| OSWorld | — | 63.3% |
| Vending-Bench 2 | — | 1,198 |
Reasoning Kimi K2.5 leads
gpt-oss-20b: 19.3 (#261), Kimi K2.5: 31.2 (#80)
| Benchmark | gpt-oss-20b | Kimi K2.5 |
|---|---|---|
| Kagi LLM Benchmark | 53.2% | 78.5% |
| CritPt | 1.4% | 3.1% |
| Chess Puzzles | 4% | 12% |
| LMArena Hard Prompts | 1274 | 1453 |
| Epoch Capabilities Index | 137.82 | 148.03 |
| ARC-AGI-2 | — | 11.8% |
| SimpleBench | — | 46.8% |
| NYT Connections (extended) | — | 69.9% |
| ARC-AGI-1 | — | 65.3% |
| EnigmaEval | — | 3.4% |
| Thematic Generalization | — | 69.4% |
| DTBench | 68% | — |
| LMCA | 14.5% | — |
Math Kimi K2.5 leads
gpt-oss-20b: 39.4 (#103), Kimi K2.5: 51.8 (#53)
| Benchmark | gpt-oss-20b | Kimi K2.5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 65.3% | 92.2% |
| LMArena Math | 1317 | 1470 |
| MathArena Final-Answer Competitions | — | 62.3% |
| Omni-MATH | 56.5% | — |
| FrontierMath (Feb 2025 set) | — | 27.9% |
| FrontierMath Tier 4 (v1) | — | 4.2% |
Knowledge Kimi K2.5 leads
gpt-oss-20b: 34.6 (#195), Kimi K2.5: 53.6 (#56)
| Benchmark | gpt-oss-20b | Kimi K2.5 |
|---|---|---|
| GPQA Diamond | 60.8% | 87.6% |
| LMArena Expert | 1258 | 1466 |
| Humanity's Last Exam | — | 24.4% |
| SimpleQA Verified | — | 34.3% |
| MMLU-Pro | 74% | — |
| Vectara Hallucination Rate | — | 14.2% |
| GPQA (HELM) | 59.4% | — |
Multimodal Not comparable
gpt-oss-20b: —, Kimi K2.5: 41.1 (#39)
| Benchmark | gpt-oss-20b | Kimi K2.5 |
|---|---|---|
| LMArena Vision | — | 1269 |
| LMArena Document | — | 1430 |
Multilingual Kimi K2.5 leads
gpt-oss-20b: 42.2 (#197), Kimi K2.5: 53.9 (#53)
| Benchmark | gpt-oss-20b | Kimi K2.5 |
|---|---|---|
| LMArena Non-English | 1268 | 1433 |
| LMArena Chinese | 1314 | 1495 |
| LMArena German | 1255 | 1441 |
| LMArena Japanese | 1244 | 1421 |
| LMArena Korean | 1236 | 1410 |
| LMArena Russian | 1278 | 1435 |
| LMArena Spanish | 1267 | 1450 |
| LMArena French | — | 1454 |
Instruction Following Kimi K2.5 leads
gpt-oss-20b: 61.8 (#240), Kimi K2.5: 75.3 (#64)
| Benchmark | gpt-oss-20b | Kimi K2.5 |
|---|---|---|
| LMArena Instruction Following | 1236 | 1431 |
| IFEval | 73.2% | — |
Long Context Kimi K2.5 leads
gpt-oss-20b: 37.9 (#209), Kimi K2.5: 52.1 (#7)
| Benchmark | gpt-oss-20b | Kimi K2.5 |
|---|---|---|
| LMArena Longer Query | 1250 | 1445 |
| Fiction.LiveBench | — | 86.1% |
| CL-bench | — | 19.3% |
| CL-bench Life | — | 13.2% |
Writing & Preference Kimi K2.5 leads
gpt-oss-20b: 35.5 (#265), Kimi K2.5: 65.1 (#53)
| Benchmark | gpt-oss-20b | Kimi K2.5 |
|---|---|---|
| LMArena Text | 1287 | 1445 |
| LMArena Creative Writing | 1201 | 1423 |
| EQ-Bench Creative Writing | 666 | 1579 |
| LMArena Multi-Turn | 1268 | 1444 |
| WildBench | 73.7% | — |
Frequently asked questions
Is gpt-oss-20b better than Kimi K2.5?
Kimi K2.5 is the stronger model overall, scoring 48.1 to 32.5 on the Noometry Index. gpt-oss-20b costs 25× less per token, which makes it the better buy when Kimi K2.5's lead doesn't matter for your workload.
Which is cheaper, gpt-oss-20b or Kimi K2.5?
gpt-oss-20b is cheaper. It lists at $0.018 per million input tokens and $0.09 per million output tokens; Kimi K2.5 lists at $0.45 and $2.25.
Is gpt-oss-20b or Kimi K2.5 better for coding?
Kimi K2.5 scores higher on coding benchmarks: 48.8 versus 37.6 in the Noometry coding category.
Which has the bigger context window?
Kimi K2.5 does, with 262K tokens against 131K.
How many benchmarks do gpt-oss-20b and Kimi K2.5 share?
27 benchmarks have published results for both models. gpt-oss-20b has 34 scored results on Noometry and Kimi K2.5 has 51.