Model comparison
GPT-4 Turbo vs Kimi K3
Kimi K3 is the stronger model overall, scoring 59.5 to 30.5 on the Noometry Index.
Last verified . 27 shared benchmarks.
Summary
- They share 27 benchmarks with published results for both. GPT-4 Turbo scores higher in 0 categories and Kimi K3 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K3 leads 74.2 to 9.0.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 6.7% for GPT-4 Turbo and 97.2% for Kimi K3.
- Kimi K3 is cheaper at $3 / $15 per million input/output tokens, against $10 / $30 for GPT-4 Turbo.
- Kimi K3 accepts more context: 1.05M tokens versus 128K.
- Kimi K3 has downloadable open weights; the other is API-only.
Side by side
| GPT-4 Turbo | Kimi K3 | |
|---|---|---|
| Provider | OpenAI | Moonshot AI |
| Noometry Index | 30.5 | 59.5 |
| Released | 2023-11-06 | 2026-07-16 |
| Weights | Proprietary | Open |
| Context window | 128K | 1.05M |
| Max output | 4K | 1.05M |
| Input $ / M tokens | $10 | $3 |
| Output $ / M tokens | $30 | $15 |
| Results tracked | 36 | 53 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K3 leads
GPT-4 Turbo: 33.8 (#249), Kimi K3: 61.0 (#10)
| Benchmark | GPT-4 Turbo | Kimi K3 |
|---|---|---|
| WeirdML | 18% | 82.6% |
| LMArena Coding | 1268 | 1508 |
| DeepSWE | — | 68.5% |
| FrontierCode | — | 44.2% |
| LMArena WebDev | — | 1654 |
| FrontierSWE | — | 25.9% |
| SciCode | — | 59.5% |
| BigCodeBench Instruct | 48.2% | — |
| BigCodeBench Complete | 58.2% | — |
| ALE-Bench | — | 1,524 |
| HumanEval+ | 86.6% | — |
| MBPP+ | 73.3% | — |
Agentic & Tool Use Not comparable
GPT-4 Turbo: —, Kimi K3: 41.8 (#20)
| Benchmark | GPT-4 Turbo | Kimi K3 |
|---|---|---|
| APEX-Agents | — | 50.6% |
| τ²-bench Banking | — | 37.1% |
| PostTrainBench | — | 32% |
| GBAEval | — | 48.3% |
| GDP.pdf | — | 19% |
| METR Time Horizons | 36.7% | — |
| Vending-Bench 2 | — | 5,165 |
Reasoning Kimi K3 leads
GPT-4 Turbo: 15.3 (#317), Kimi K3: 63.0 (#17)
| Benchmark | GPT-4 Turbo | Kimi K3 |
|---|---|---|
| SimpleBench | 25.1% | 60.7% |
| Chess Puzzles | 6% | 39% |
| LMArena Hard Prompts | 1251 | 1496 |
| DTBench | 61.6% | 91.2% |
| LMCA | 9.8% | 52.7% |
| Epoch Capabilities Index | 127.25 | 157.45 |
| ForecastBench | 59.4 | 61.1 |
| ARC-AGI-2 | — | 60.4% |
| NYT Connections (extended) | — | 93.6% |
| ARC-AGI-1 | — | 94.5% |
| CritPt | — | 23.4% |
| Mystery Game Puzzles | — | 26% |
| Surface Evolver Bench | — | 95% |
Math Kimi K3 leads
GPT-4 Turbo: 9.0 (#322), Kimi K3: 74.2 (#16)
| Benchmark | GPT-4 Turbo | Kimi K3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 0.7% | 72.2% |
| OTIS Mock AIME 2024-2025 | 6.7% | 97.2% |
| LMArena Math | 1272 | 1491 |
| FrontierMath Tier 4 | — | 39% |
| MathArena Final-Answer Competitions | — | 87.8% |
| ProofBench | — | 87% |
| MATH Level 5 | 46.7% | — |
Knowledge Kimi K3 leads
GPT-4 Turbo: 24.3 (#268), Kimi K3: 63.2 (#21)
| Benchmark | GPT-4 Turbo | Kimi K3 |
|---|---|---|
| GPQA Diamond | 46.6% | 93.1% |
| LMArena Expert | 1223 | 1521 |
| SimpleQA Verified | — | 50.6% |
| Confabulations | 28.4% | — |
| MMLU | 81.3% | — |
Multimodal Kimi K3 leads
GPT-4 Turbo: 30.6 (#110), Kimi K3: 37.8 (#70)
| Benchmark | GPT-4 Turbo | Kimi K3 |
|---|---|---|
| LMArena Vision | 1090 | — |
| Blueprint-Bench 2 | — | 29.5% |
| Furniture Assembly | — | 34.2% |
Multilingual Kimi K3 leads
GPT-4 Turbo: 40.5 (#216), Kimi K3: 56.3 (#21)
| Benchmark | GPT-4 Turbo | Kimi K3 |
|---|---|---|
| LMArena Non-English | 1245 | 1466 |
| LMArena Chinese | 1242 | 1529 |
| LMArena French | 1276 | 1491 |
| LMArena German | 1259 | 1488 |
| LMArena Japanese | 1194 | 1487 |
| LMArena Korean | 1187 | 1458 |
| LMArena Russian | 1259 | 1482 |
| LMArena Spanish | 1260 | 1472 |
Instruction Following Kimi K3 leads
GPT-4 Turbo: 65.8 (#216), Kimi K3: 77.7 (#14)
| Benchmark | GPT-4 Turbo | Kimi K3 |
|---|---|---|
| LMArena Instruction Following | 1249 | 1483 |
Long Context Kimi K3 leads
GPT-4 Turbo: 38.0 (#206), Kimi K3: 45.8 (#29)
| Benchmark | GPT-4 Turbo | Kimi K3 |
|---|---|---|
| LMArena Longer Query | 1254 | 1494 |
Writing & Preference Kimi K3 leads
GPT-4 Turbo: 47.7 (#206), Kimi K3: 76.6 (#4)
| Benchmark | GPT-4 Turbo | Kimi K3 |
|---|---|---|
| LMArena Text | 1272 | 1476 |
| LMArena Creative Writing | 1269 | 1454 |
| LMArena Multi-Turn | 1267 | 1488 |
| EQ-Bench Creative Writing | — | 2082 |
| EQ-Bench 4 | — | 1339 |
Frequently asked questions
Is GPT-4 Turbo better than Kimi K3?
Kimi K3 is the stronger model overall, scoring 59.5 to 30.5 on the Noometry Index.
Which is cheaper, GPT-4 Turbo or Kimi K3?
Kimi K3 is cheaper. It lists at $3 per million input tokens and $15 per million output tokens; GPT-4 Turbo lists at $10 and $30.
Is GPT-4 Turbo or Kimi K3 better for coding?
Kimi K3 scores higher on coding benchmarks: 61.0 versus 33.8 in the Noometry coding category.
Which has the bigger context window?
Kimi K3 does, with 1.05M tokens against 128K.
How many benchmarks do GPT-4 Turbo and Kimi K3 share?
27 benchmarks have published results for both models. GPT-4 Turbo has 36 scored results on Noometry and Kimi K3 has 53.