Model comparison
GPT-4 vs Kimi K3
Kimi K3 is the stronger model overall, scoring 59.5 to 29.1 on the Noometry Index.
Last verified . 27 shared benchmarks.
Summary
- They share 27 benchmarks with published results for both. GPT-4 scores higher in 0 categories and Kimi K3 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K3 leads 74.2 to 10.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 1.1% for GPT-4 and 97.2% for Kimi K3.
- Kimi K3 is cheaper at $3 / $15 per million input/output tokens, against $30 / $60 for GPT-4.
- Kimi K3 accepts more context: 1.05M tokens versus 8K.
- Kimi K3 has downloadable open weights; the other is API-only.
Side by side
| GPT-4 | Kimi K3 | |
|---|---|---|
| Provider | OpenAI | Moonshot AI |
| Noometry Index | 29.1 | 59.5 |
| Released | 2023-03-14 | 2026-07-16 |
| Weights | Proprietary | Open |
| Context window | 8K | 1.05M |
| Max output | 8K | 1.05M |
| Input $ / M tokens | $30 | $3 |
| Output $ / M tokens | $60 | $15 |
| Results tracked | 38 | 53 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K3 leads
GPT-4: 31.6 (#283), Kimi K3: 61.0 (#10)
| Benchmark | GPT-4 | Kimi K3 |
|---|---|---|
| WeirdML | 12.4% | 82.6% |
| LMArena Coding | 1254 | 1508 |
| DeepSWE | — | 68.5% |
| FrontierCode | — | 44.2% |
| LMArena WebDev | — | 1654 |
| FrontierSWE | — | 25.9% |
| SciCode | — | 59.5% |
| BigCodeBench Instruct | 46% | — |
| BigCodeBench Complete | 57.2% | — |
| ALE-Bench | — | 1,524 |
| HumanEval+ | 79.3% | — |
Agentic & Tool Use Not comparable
GPT-4: —, Kimi K3: 41.8 (#20)
| Benchmark | GPT-4 | Kimi K3 |
|---|---|---|
| APEX-Agents | — | 50.6% |
| τ²-bench Banking | — | 37.1% |
| PostTrainBench | — | 32% |
| GBAEval | — | 48.3% |
| GDP.pdf | — | 19% |
| METR Time Horizons | 36.1% | — |
| Vending-Bench 2 | — | 5,165 |
Reasoning Kimi K3 leads
GPT-4: 17.8 (#289), Kimi K3: 63.0 (#17)
| Benchmark | GPT-4 | Kimi K3 |
|---|---|---|
| Chess Puzzles | 4% | 39% |
| LMArena Hard Prompts | 1241 | 1496 |
| Mystery Game Puzzles | 12% | 26% |
| DTBench | 62.7% | 91.2% |
| LMCA | 17.1% | 52.7% |
| Epoch Capabilities Index | 125.89 | 157.45 |
| ForecastBench | 57.8 | 61.1 |
| ARC-AGI-2 | — | 60.4% |
| SimpleBench | — | 60.7% |
| NYT Connections (extended) | — | 93.6% |
| ARC-AGI-1 | — | 94.5% |
| CritPt | — | 23.4% |
| Surface Evolver Bench | — | 95% |
| BIG-Bench Hard | 75.1% | — |
| HellaSwag | 95.3% | — |
| WinoGrande | 87.5% | — |
Math Kimi K3 leads
GPT-4: 10.8 (#309), Kimi K3: 74.2 (#16)
| Benchmark | GPT-4 | Kimi K3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.1% | 97.2% |
| LMArena Math | 1269 | 1491 |
| FrontierMath (Tiers 1-3) | — | 72.2% |
| FrontierMath Tier 4 | — | 39% |
| MathArena Final-Answer Competitions | — | 87.8% |
| ProofBench | — | 87% |
| MATH Level 5 | 23% | — |
| GSM8K | 92% | — |
Knowledge Kimi K3 leads
GPT-4: 18.4 (#282), Kimi K3: 63.2 (#21)
| Benchmark | GPT-4 | Kimi K3 |
|---|---|---|
| GPQA Diamond | 35.7% | 93.1% |
| LMArena Expert | 1211 | 1521 |
| SimpleQA Verified | — | 50.6% |
| MMLU | 86.4% | — |
| TriviaQA | 84.8% | — |
Multimodal Not comparable
GPT-4: —, Kimi K3: 37.8 (#70)
| Benchmark | GPT-4 | Kimi K3 |
|---|---|---|
| Blueprint-Bench 2 | — | 29.5% |
| Furniture Assembly | — | 34.2% |
Multilingual Kimi K3 leads
GPT-4: 40.6 (#215), Kimi K3: 56.3 (#21)
| Benchmark | GPT-4 | Kimi K3 |
|---|---|---|
| LMArena Non-English | 1246 | 1466 |
| LMArena Chinese | 1242 | 1529 |
| LMArena French | 1283 | 1491 |
| LMArena German | 1251 | 1488 |
| LMArena Japanese | 1209 | 1487 |
| LMArena Korean | 1184 | 1458 |
| LMArena Russian | 1251 | 1482 |
| LMArena Spanish | 1261 | 1472 |
Instruction Following Kimi K3 leads
GPT-4: 65.3 (#222), Kimi K3: 77.7 (#14)
| Benchmark | GPT-4 | Kimi K3 |
|---|---|---|
| LMArena Instruction Following | 1241 | 1483 |
Long Context Kimi K3 leads
GPT-4: 37.7 (#212), Kimi K3: 45.8 (#29)
| Benchmark | GPT-4 | Kimi K3 |
|---|---|---|
| LMArena Longer Query | 1244 | 1494 |
Writing & Preference Kimi K3 leads
GPT-4: 34.9 (#268), Kimi K3: 76.6 (#4)
| Benchmark | GPT-4 | Kimi K3 |
|---|---|---|
| LMArena Text | 1263 | 1476 |
| LMArena Creative Writing | 1244 | 1454 |
| EQ-Bench Creative Writing | 752 | 2082 |
| LMArena Multi-Turn | 1257 | 1488 |
| EQ-Bench 4 | — | 1339 |
Frequently asked questions
Is GPT-4 better than Kimi K3?
Kimi K3 is the stronger model overall, scoring 59.5 to 29.1 on the Noometry Index.
Which is cheaper, GPT-4 or Kimi K3?
Kimi K3 is cheaper. It lists at $3 per million input tokens and $15 per million output tokens; GPT-4 lists at $30 and $60.
Is GPT-4 or Kimi K3 better for coding?
Kimi K3 scores higher on coding benchmarks: 61.0 versus 31.6 in the Noometry coding category.
Which has the bigger context window?
Kimi K3 does, with 1.05M tokens against 8K.
How many benchmarks do GPT-4 and Kimi K3 share?
27 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and Kimi K3 has 53.