Model comparison
Kimi K2.5 vs o3-mini
Kimi K2.5 is the stronger model overall, scoring 48.1 to 36.7 on the Noometry Index.
Last verified . 31 shared benchmarks.
Summary
- They share 31 benchmarks with published results for both. Kimi K2.5 scores higher in 9 categories and o3-mini in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K2.5 leads 51.8 to 28.1.
- The biggest single-benchmark swing is Fiction.LiveBench: 86.1% for Kimi K2.5 and 50% for o3-mini.
- Kimi K2.5 is cheaper at $0.45 / $2.25 per million input/output tokens, against $1.10 / $4.40 for o3-mini.
- Kimi K2.5 accepts more context: 262K tokens versus 200K.
- Kimi K2.5 has downloadable open weights; the other is API-only.
Side by side
| Kimi K2.5 | o3-mini | |
|---|---|---|
| Provider | Moonshot AI | OpenAI |
| Noometry Index | 48.1 | 36.7 |
| Released | 2026-01-27 | 2024-12-20 |
| Weights | Open | Proprietary |
| Context window | 262K | 200K |
| Max output | 262K | 100K |
| Input $ / M tokens | $0.45 | $1.10 |
| Output $ / M tokens | $2.25 | $4.40 |
| Results tracked | 51 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K2.5 leads
Kimi K2.5: 48.8 (#53), o3-mini: 40.8 (#132)
| Benchmark | Kimi K2.5 | o3-mini |
|---|---|---|
| SciCode | 49% | 39.8% |
| WeirdML | 45.6% | 43.7% |
| LMArena Coding | 1474 | 1378 |
| SWE-bench Verified | 73.8% | — |
| SWE-bench Verified (bash only) | 70.8% | — |
| Aider Polyglot | — | 60.4% |
| LMArena WebDev | 1437 | — |
| SWE-bench Multilingual | 67.3% | — |
| GSO | — | 1.3% |
| LiveBench Coding | — | 82.7% |
| CadEval | — | 54% |
| ALE-Bench | 821.65 | — |
Agentic & Tool Use Kimi K2.5 leads
Kimi K2.5: 34.2 (#48), o3-mini: 29.6 (#84)
| Benchmark | Kimi K2.5 | o3-mini |
|---|---|---|
| Terminal-Bench | 43.2% | — |
| Cybench | — | 22.5% |
| OSWorld | 63.3% | — |
| Vending-Bench 2 | 1,198 | — |
Reasoning Kimi K2.5 leads
Kimi K2.5: 31.2 (#80), o3-mini: 16.3 (#305)
| Benchmark | Kimi K2.5 | o3-mini |
|---|---|---|
| ARC-AGI-2 | 11.8% | 3% |
| SimpleBench | 46.8% | 22.8% |
| ARC-AGI-1 | 65.3% | 34.5% |
| CritPt | 3.1% | 0.3% |
| Chess Puzzles | 12% | 17% |
| LMArena Hard Prompts | 1453 | 1366 |
| Epoch Capabilities Index | 148.03 | 140.34 |
| Kagi LLM Benchmark | 78.5% | — |
| NYT Connections (extended) | 69.9% | — |
| EnigmaEval | 3.4% | — |
| Thematic Generalization | 69.4% | — |
| LiveBench Reasoning | — | 89.6% |
| Mystery Game Puzzles | — | 7% |
| DTBench | — | 68.8% |
| LiveBench Data Analysis | — | 70.6% |
| LMCA | — | 19% |
| ForecastBench | — | 59.6 |
| LiveBench | — | 75.9% |
Math Kimi K2.5 leads
Kimi K2.5: 51.8 (#53), o3-mini: 28.1 (#244)
| Benchmark | Kimi K2.5 | o3-mini |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 92.2% | 76.9% |
| LMArena Math | 1470 | 1396 |
| FrontierMath (Feb 2025 set) | 27.9% | 12.4% |
| FrontierMath Tier 4 (v1) | 4.2% | 4.2% |
| FrontierMath (Tiers 1-3) | — | 18.6% |
| FrontierMath Tier 4 | — | 0% |
| MathArena Final-Answer Competitions | 62.3% | — |
| LiveBench Math | — | 77.3% |
| MATH Level 5 | — | 96.5% |
Knowledge Kimi K2.5 leads
Kimi K2.5: 53.6 (#56), o3-mini: 38.3 (#146)
| Benchmark | Kimi K2.5 | o3-mini |
|---|---|---|
| GPQA Diamond | 87.6% | 77% |
| SimpleQA Verified | 34.3% | 15.3% |
| LMArena Expert | 1466 | 1364 |
| Humanity's Last Exam | 24.4% | — |
| Confabulations | — | 17.9% |
| Vectara Hallucination Rate | 14.2% | — |
Multimodal Not comparable
Kimi K2.5: 41.1 (#39), o3-mini: —
| Benchmark | Kimi K2.5 | o3-mini |
|---|---|---|
| LMArena Vision | 1269 | — |
| LMArena Document | 1430 | — |
Multilingual Kimi K2.5 leads
Kimi K2.5: 53.9 (#53), o3-mini: 45.7 (#164)
| Benchmark | Kimi K2.5 | o3-mini |
|---|---|---|
| LMArena Non-English | 1433 | 1319 |
| LMArena Chinese | 1495 | 1379 |
| LMArena French | 1454 | 1334 |
| LMArena German | 1441 | 1303 |
| LMArena Japanese | 1421 | 1286 |
| LMArena Korean | 1410 | 1314 |
| LMArena Russian | 1435 | 1304 |
| LMArena Spanish | 1450 | 1321 |
Instruction Following Too close to call
Kimi K2.5: 75.3 (#64), o3-mini: 75.1 (#72)
| Benchmark | Kimi K2.5 | o3-mini |
|---|---|---|
| LMArena Instruction Following | 1431 | 1337 |
| LiveBench Instruction Following | — | 84.4% |
Long Context Kimi K2.5 leads
Kimi K2.5: 52.1 (#7), o3-mini: 33.8 (#256)
| Benchmark | Kimi K2.5 | o3-mini |
|---|---|---|
| Fiction.LiveBench | 86.1% | 50% |
| LMArena Longer Query | 1445 | 1343 |
| CL-bench | 19.3% | — |
| CL-bench Life | 13.2% | — |
Writing & Preference Kimi K2.5 leads
Kimi K2.5: 65.1 (#53), o3-mini: 50.3 (#182)
| Benchmark | Kimi K2.5 | o3-mini |
|---|---|---|
| LMArena Text | 1445 | 1337 |
| LMArena Creative Writing | 1423 | 1286 |
| LMArena Multi-Turn | 1444 | 1320 |
| Short-Story Creative Writing | — | 61.7% |
| EQ-Bench Creative Writing | 1579 | — |
| LiveBench Language | — | 50.7% |
Frequently asked questions
Is Kimi K2.5 better than o3-mini?
Kimi K2.5 is the stronger model overall, scoring 48.1 to 36.7 on the Noometry Index.
Which is cheaper, Kimi K2.5 or o3-mini?
Kimi K2.5 is cheaper. It lists at $0.45 per million input tokens and $2.25 per million output tokens; o3-mini lists at $1.10 and $4.40.
Is Kimi K2.5 or o3-mini better for coding?
Kimi K2.5 scores higher on coding benchmarks: 48.8 versus 40.8 in the Noometry coding category.
Which has the bigger context window?
Kimi K2.5 does, with 262K tokens against 200K.
How many benchmarks do Kimi K2.5 and o3-mini share?
31 benchmarks have published results for both models. Kimi K2.5 has 51 scored results on Noometry and o3-mini has 51.