Model comparison
GPT-5.2 vs Kimi K2.6
GPT-5.2 is the stronger model overall, scoring 54.1 to 47.7 on the Noometry Index. Kimi K2.6 costs 2.8× less per token, which makes it the better buy when GPT-5.2's lead doesn't matter for your workload.
Last verified . 43 shared benchmarks.
Summary
- They share 43 benchmarks with published results for both. GPT-5.2 scores higher in 6 categories and Kimi K2.6 in 4 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in multimodal, where GPT-5.2 leads 51.3 to 31.6.
- The biggest single-benchmark swing is Chess Puzzles: 49% for GPT-5.2 and 26% for Kimi K2.6.
- Kimi K2.6 is cheaper at $0.95 / $4 per million input/output tokens, against $1.75 / $14 for GPT-5.2.
- GPT-5.2 accepts more context: 400K tokens versus 262K.
- Kimi K2.6 has downloadable open weights; the other is API-only.
Side by side
| GPT-5.2 | Kimi K2.6 | |
|---|---|---|
| Provider | OpenAI | Moonshot AI |
| Noometry Index | 54.1 | 47.7 |
| Released | 2025-12-11 | 2026-04-20 |
| Weights | Proprietary | Open |
| Context window | 400K | 262K |
| Max output | 128K | 262K |
| Input $ / M tokens | $1.75 | $0.95 |
| Output $ / M tokens | $14 | $4 |
| Results tracked | 67 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
GPT-5.2: 51.6 (#37), Kimi K2.6: 50.7 (#43)
| Benchmark | GPT-5.2 | Kimi K2.6 |
|---|---|---|
| SWE-bench Verified | 73.8% | 76.7% |
| LMArena WebDev | 1416 | 1509 |
| WeirdML | 72.2% | 55.9% |
| LMArena Coding | 1447 | 1488 |
| ALE-Bench | 1,294 | 1,093 |
| SWE-bench Verified (bash only) | 72.8% | — |
| SWE-bench Multilingual | 66.7% | — |
| SciCode | — | 53.5% |
| GSO | 27.4% | — |
| AlgoTune | 2.05 | — |
Agentic & Tool Use GPT-5.2 leads
GPT-5.2: 40.2 (#24), Kimi K2.6: 21.9 (#137)
| Benchmark | GPT-5.2 | Kimi K2.6 |
|---|---|---|
| Vending-Bench 2 | 3,591 | 6,205 |
| Terminal-Bench | 64.9% | — |
| Berkeley Function Calling Leaderboard | 55.9% | — |
| OSWorld 2.0 | — | 4.6% |
| GDPval | 49.7% | — |
| Remote Labor Index | 2.5% | — |
| τ²-bench Airline | 83% | — |
| τ²-bench Banking | 32.2% | — |
| τ²-bench Retail | 81.6% | — |
| τ²-bench Telecom | 89.7% | — |
| DeepResearch Bench | 41.1% | — |
| ExploitBench | — | 18.4% |
| GBAEval | — | 0.9% |
| GDP.pdf | — | 12% |
| LMArena Search | 1207 | — |
| METR Time Horizons | 75.3% | — |
Reasoning GPT-5.2 leads
GPT-5.2: 50.2 (#35), Kimi K2.6: 40.5 (#55)
| Benchmark | GPT-5.2 | Kimi K2.6 |
|---|---|---|
| NYT Connections (extended) | 83.6% | 87.2% |
| Chess Puzzles | 49% | 26% |
| EBR-Bench | 23% | 2.4% |
| LMArena Hard Prompts | 1445 | 1470 |
| Mystery Game Puzzles | 23% | 18% |
| DTBench | 90.9% | 90.9% |
| LMCA | 43.9% | 37.3% |
| Epoch Capabilities Index | 153.45 | 151.05 |
| ARC-AGI-2 | 52.9% | — |
| SimpleBench | 45.8% | — |
| Kagi LLM Benchmark | 73.3% | — |
| ARC-AGI-1 | 86.2% | — |
| CritPt | — | 8% |
| EnigmaEval | 10.4% | — |
| ForecastBench | 60.1 | — |
Math GPT-5.2 leads
GPT-5.2: 60.0 (#38), Kimi K2.6: 57.0 (#41)
| Benchmark | GPT-5.2 | Kimi K2.6 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 67.4% | 57.2% |
| FrontierMath Tier 4 | 31.7% | 25.6% |
| MathArena Final-Answer Competitions | 72% | 72.9% |
| OTIS Mock AIME 2024-2025 | 96.1% | 96.1% |
| ProofBench | 15% | 16% |
| LMArena Math | 1440 | 1475 |
| FrontierMath (Feb 2025 set) | 40.7% | 39% |
| FrontierMath Tier 4 (v1) | 18.8% | 14.6% |
Knowledge GPT-5.2 leads
GPT-5.2: 59.3 (#32), Kimi K2.6: 54.0 (#54)
| Benchmark | GPT-5.2 | Kimi K2.6 |
|---|---|---|
| GPQA Diamond | 91.4% | 90.8% |
| SimpleQA Verified | 37.1% | 34.9% |
| Vectara Hallucination Rate | 8.4% | 10.8% |
| LMArena Expert | 1445 | 1491 |
| Humanity's Last Exam | 27.8% | — |
Multimodal GPT-5.2 leads
GPT-5.2: 51.3 (#7), Kimi K2.6: 31.6 (#103)
| Benchmark | GPT-5.2 | Kimi K2.6 |
|---|---|---|
| LMArena Vision | 1268 | 1283 |
| Furniture Assembly | 38.3% | 21.7% |
| LMArena Document | 1405 | 1451 |
| VPCT | 84% | — |
| Blueprint-Bench 2 | — | 3.9% |
Multilingual Kimi K2.6 leads
GPT-5.2: 53.4 (#67), Kimi K2.6: 54.9 (#37)
| Benchmark | GPT-5.2 | Kimi K2.6 |
|---|---|---|
| LMArena Non-English | 1425 | 1446 |
| LMArena Chinese | 1460 | 1521 |
| LMArena French | 1455 | 1471 |
| LMArena German | 1448 | 1450 |
| LMArena Japanese | 1420 | 1443 |
| LMArena Korean | 1392 | 1427 |
| LMArena Russian | 1440 | 1446 |
| LMArena Spanish | 1433 | 1464 |
Instruction Following Kimi K2.6 leads
GPT-5.2: 74.7 (#89), Kimi K2.6: 76.3 (#43)
| Benchmark | GPT-5.2 | Kimi K2.6 |
|---|---|---|
| LMArena Instruction Following | 1417 | 1451 |
Long Context Too close to call
GPT-5.2: 44.0 (#78), Kimi K2.6: 44.9 (#52)
| Benchmark | GPT-5.2 | Kimi K2.6 |
|---|---|---|
| LMArena Longer Query | 1428 | 1468 |
| CL-bench | 18.2% | — |
Writing & Preference Kimi K2.6 leads
GPT-5.2: 66.8 (#32), Kimi K2.6: 68.5 (#26)
| Benchmark | GPT-5.2 | Kimi K2.6 |
|---|---|---|
| LMArena Text | 1439 | 1455 |
| LMArena Creative Writing | 1401 | 1434 |
| EQ-Bench Creative Writing | 1703 | 1725 |
| LMArena Multi-Turn | 1458 | 1453 |
| EQ-Bench 4 | — | 1202 |
Frequently asked questions
Is GPT-5.2 better than Kimi K2.6?
GPT-5.2 is the stronger model overall, scoring 54.1 to 47.7 on the Noometry Index. Kimi K2.6 costs 2.8× less per token, which makes it the better buy when GPT-5.2's lead doesn't matter for your workload.
Which is cheaper, GPT-5.2 or Kimi K2.6?
Kimi K2.6 is cheaper. It lists at $0.95 per million input tokens and $4 per million output tokens; GPT-5.2 lists at $1.75 and $14.
Is GPT-5.2 or Kimi K2.6 better for coding?
They score almost the same on coding (51.6 vs 50.7); test both on your own repository before choosing.
Which has the bigger context window?
GPT-5.2 does, with 400K tokens against 262K.
How many benchmarks do GPT-5.2 and Kimi K2.6 share?
43 benchmarks have published results for both models. GPT-5.2 has 67 scored results on Noometry and Kimi K2.6 has 51.