Model comparison
GPT-4 Turbo vs Kimi K2.6
Kimi K2.6 is the stronger model overall, scoring 47.5 to 30.3 on the Noometry Index.
Last verified . 26 shared benchmarks.
Summary
- They share 26 benchmarks with published results for both. GPT-4 Turbo scores higher in 0 categories and Kimi K2.6 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K2.6 leads 56.8 to 8.6.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 6.7% for GPT-4 Turbo and 96.1% for Kimi K2.6.
- Kimi K2.6 is cheaper at $0.95 / $4 per million input/output tokens, against $10 / $30 for GPT-4 Turbo.
- Kimi K2.6 accepts more context: 262K tokens versus 128K.
- Kimi K2.6 has downloadable open weights; the other is API-only.
Side by side
| GPT-4 Turbo | Kimi K2.6 | |
|---|---|---|
| Provider | OpenAI | Moonshot AI |
| Noometry Index | 30.3 | 47.5 |
| Released | 2023-11-06 | 2026-04-20 |
| Weights | Proprietary | Open |
| Context window | 128K | 262K |
| Max output | 4K | 262K |
| Input $ / M tokens | $10 | $0.95 |
| Output $ / M tokens | $30 | $4 |
| Results tracked | 36 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K2.6 leads
GPT-4 Turbo: 33.4 (#247), Kimi K2.6: 50.5 (#41)
| Benchmark | GPT-4 Turbo | Kimi K2.6 |
|---|---|---|
| WeirdML | 18% | 55.9% |
| LMArena Coding | 1268 | 1488 |
| SWE-bench Verified | — | 76.7% |
| LMArena WebDev | — | 1509 |
| SciCode | — | 53.5% |
| BigCodeBench Instruct | 48.2% | — |
| BigCodeBench Complete | 58.2% | — |
| ALE-Bench | — | 1,093 |
| HumanEval+ | 86.6% | — |
| MBPP+ | 73.3% | — |
Agentic & Tool Use Not comparable
GPT-4 Turbo: —, Kimi K2.6: 21.8 (#128)
| Benchmark | GPT-4 Turbo | Kimi K2.6 |
|---|---|---|
| OSWorld 2.0 | — | 4.6% |
| ExploitBench | — | 18.4% |
| GBAEval | — | 0.9% |
| GDP.pdf | — | 12% |
| METR Time Horizons | 36.7% | — |
| Vending-Bench 2 | — | 6,205 |
Reasoning Kimi K2.6 leads
GPT-4 Turbo: 14.9 (#311), Kimi K2.6: 39.6 (#51)
| Benchmark | GPT-4 Turbo | Kimi K2.6 |
|---|---|---|
| Chess Puzzles | 6% | 26% |
| LMArena Hard Prompts | 1251 | 1470 |
| DTBench | 61.6% | 90.9% |
| LMCA | 9.8% | 37.3% |
| Epoch Capabilities Index | 127.25 | 151.03 |
| SimpleBench | 25.1% | — |
| NYT Connections (extended) | — | 87.2% |
| CritPt | — | 8% |
| EBR-Bench | — | 2.4% |
| Mystery Game Puzzles | — | 18% |
| ForecastBench | 59.4 | — |
Math Kimi K2.6 leads
GPT-4 Turbo: 8.6 (#316), Kimi K2.6: 56.8 (#34)
| Benchmark | GPT-4 Turbo | Kimi K2.6 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 0.7% | 57.2% |
| OTIS Mock AIME 2024-2025 | 6.7% | 96.1% |
| LMArena Math | 1272 | 1475 |
| FrontierMath Tier 4 | — | 25.6% |
| MathArena Final-Answer Competitions | — | 72.9% |
| ProofBench | — | 16% |
| MATH Level 5 | 46.7% | — |
| FrontierMath (Feb 2025 set) | — | 39% |
| FrontierMath Tier 4 (v1) | — | 14.6% |
Knowledge Kimi K2.6 leads
GPT-4 Turbo: 24.3 (#263), Kimi K2.6: 53.9 (#48)
| Benchmark | GPT-4 Turbo | Kimi K2.6 |
|---|---|---|
| GPQA Diamond | 46.6% | 90.8% |
| LMArena Expert | 1223 | 1491 |
| SimpleQA Verified | — | 34.9% |
| Confabulations | 28.4% | — |
| Vectara Hallucination Rate | — | 10.8% |
| MMLU | 81.3% | — |
Multimodal Kimi K2.6 leads
GPT-4 Turbo: 30.0 (#104), Kimi K2.6: 31.8 (#95)
| Benchmark | GPT-4 Turbo | Kimi K2.6 |
|---|---|---|
| LMArena Vision | 1090 | 1282 |
| Blueprint-Bench 2 | — | 3.9% |
| Furniture Assembly | — | 21.7% |
| LMArena Document | — | 1451 |
Multilingual Kimi K2.6 leads
GPT-4 Turbo: 40.7 (#211), Kimi K2.6: 55.1 (#35)
| Benchmark | GPT-4 Turbo | Kimi K2.6 |
|---|---|---|
| LMArena Non-English | 1245 | 1446 |
| LMArena Chinese | 1242 | 1521 |
| LMArena French | 1275 | 1471 |
| LMArena German | 1259 | 1449 |
| LMArena Japanese | 1194 | 1442 |
| LMArena Korean | 1187 | 1428 |
| LMArena Russian | 1259 | 1446 |
| LMArena Spanish | 1260 | 1463 |
Instruction Following Kimi K2.6 leads
GPT-4 Turbo: 66.0 (#211), Kimi K2.6: 76.5 (#40)
| Benchmark | GPT-4 Turbo | Kimi K2.6 |
|---|---|---|
| LMArena Instruction Following | 1250 | 1451 |
Long Context Kimi K2.6 leads
GPT-4 Turbo: 38.0 (#200), Kimi K2.6: 44.9 (#50)
| Benchmark | GPT-4 Turbo | Kimi K2.6 |
|---|---|---|
| LMArena Longer Query | 1254 | 1468 |
Writing & Preference Kimi K2.6 leads
GPT-4 Turbo: 47.9 (#201), Kimi K2.6: 68.7 (#25)
| Benchmark | GPT-4 Turbo | Kimi K2.6 |
|---|---|---|
| LMArena Text | 1272 | 1455 |
| LMArena Creative Writing | 1269 | 1434 |
| LMArena Multi-Turn | 1267 | 1453 |
| EQ-Bench Creative Writing | — | 1725 |
| EQ-Bench 4 | — | 1202 |
Frequently asked questions
Is GPT-4 Turbo better than Kimi K2.6?
Kimi K2.6 is the stronger model overall, scoring 47.5 to 30.3 on the Noometry Index.
Which is cheaper, GPT-4 Turbo 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-4 Turbo lists at $10 and $30.
Is GPT-4 Turbo or Kimi K2.6 better for coding?
Kimi K2.6 scores higher on coding benchmarks: 50.5 versus 33.4 in the Noometry coding category.
Which has the bigger context window?
Kimi K2.6 does, with 262K tokens against 128K.
How many benchmarks do GPT-4 Turbo and Kimi K2.6 share?
26 benchmarks have published results for both models. GPT-4 Turbo has 36 scored results on Noometry and Kimi K2.6 has 51.