Model comparison
GPT-4o mini vs Kimi K3
Kimi K3 is the stronger model overall, scoring 59.5 to 25.5 on the Noometry Index. GPT-4o mini costs 23× less per token, which makes it the better buy when Kimi K3's lead doesn't matter for your workload.
Last verified . 30 shared benchmarks.
Summary
- They share 30 benchmarks with published results for both. GPT-4o mini scores higher in 0 categories and Kimi K3 in 10 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K3 leads 74.2 to 10.4.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 6.9% for GPT-4o mini and 97.2% for Kimi K3.
- GPT-4o mini is cheaper at $0.15 / $0.60 per million input/output tokens, against $3 / $15 for Kimi K3.
- 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-4o mini | Kimi K3 | |
|---|---|---|
| Provider | OpenAI | Moonshot AI |
| Noometry Index | 25.5 | 59.5 |
| Released | 2024-07-18 | 2026-07-16 |
| Weights | Proprietary | Open |
| Context window | 128K | 1.05M |
| Max output | 16K | 1.05M |
| Input $ / M tokens | $0.15 | $3 |
| Output $ / M tokens | $0.60 | $15 |
| Results tracked | 60 | 53 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K3 leads
GPT-4o mini: 22.0 (#335), Kimi K3: 61.0 (#10)
| Benchmark | GPT-4o mini | Kimi K3 |
|---|---|---|
| WeirdML | 11.8% | 82.6% |
| LMArena Coding | 1290 | 1508 |
| DeepSWE | — | 68.5% |
| FrontierCode | — | 44.2% |
| Aider Polyglot | 3.6% | — |
| LMArena WebDev | — | 1654 |
| FrontierSWE | — | 25.9% |
| SciCode | — | 59.5% |
| BigCodeBench Instruct | 46.1% | — |
| LiveBench Coding | 43.1% | — |
| BigCodeBench Complete | 57.4% | — |
| ALE-Bench | — | 1,524 |
| HumanEval+ | 83.5% | — |
| MBPP+ | 72.2% | — |
Agentic & Tool Use Kimi K3 leads
GPT-4o mini: 27.5 (#101), Kimi K3: 41.8 (#20)
| Benchmark | GPT-4o mini | Kimi K3 |
|---|---|---|
| APEX-Agents | — | 50.6% |
| τ²-bench Banking | — | 37.1% |
| PostTrainBench | — | 32% |
| BALROG | 17.4% | — |
| GBAEval | — | 48.3% |
| GDP.pdf | — | 19% |
| Vending-Bench 2 | — | 5,165 |
Reasoning Kimi K3 leads
GPT-4o mini: 8.7 (#347), Kimi K3: 63.0 (#17)
| Benchmark | GPT-4o mini | Kimi K3 |
|---|---|---|
| ARC-AGI-2 | 0% | 60.4% |
| SimpleBench | 10.7% | 60.7% |
| Chess Puzzles | 0% | 39% |
| LMArena Hard Prompts | 1267 | 1496 |
| Mystery Game Puzzles | 12% | 26% |
| DTBench | 54.4% | 91.2% |
| LMCA | 10.4% | 52.7% |
| Epoch Capabilities Index | 126.56 | 157.45 |
| Kagi LLM Benchmark | 28.8% | — |
| NYT Connections (extended) | — | 93.6% |
| ARC-AGI-1 | — | 94.5% |
| CritPt | — | 23.4% |
| LiveBench Reasoning | 32.8% | — |
| LiveBench Data Analysis | 50% | — |
| Surface Evolver Bench | — | 95% |
| ForecastBench | — | 61.1 |
| LiveBench | 41.3% | — |
| PIQA | 88.7% | — |
Math Kimi K3 leads
GPT-4o mini: 10.4 (#314), Kimi K3: 74.2 (#16)
| Benchmark | GPT-4o mini | Kimi K3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 0.7% | 72.2% |
| OTIS Mock AIME 2024-2025 | 6.9% | 97.2% |
| LMArena Math | 1267 | 1491 |
| FrontierMath Tier 4 | — | 39% |
| MathArena Final-Answer Competitions | — | 87.8% |
| ProofBench | — | 87% |
| Omni-MATH | 28% | — |
| LiveBench Math | 36.3% | — |
| MATH Level 5 | 52.6% | — |
| GSM8K | 91.3% | — |
Knowledge Kimi K3 leads
GPT-4o mini: 17.7 (#284), Kimi K3: 63.2 (#21)
| Benchmark | GPT-4o mini | Kimi K3 |
|---|---|---|
| GPQA Diamond | 37.7% | 93.1% |
| SimpleQA Verified | 8.3% | 50.6% |
| LMArena Expert | 1235 | 1521 |
| MMLU-Pro | 60.3% | — |
| Confabulations | 37.2% | — |
| GPQA (HELM) | 36.8% | — |
| BoolQ | 88.7% | — |
| MMLU | 81.8% | — |
Multimodal Kimi K3 leads
GPT-4o mini: 25.9 (#122), Kimi K3: 37.8 (#70)
| Benchmark | GPT-4o mini | Kimi K3 |
|---|---|---|
| LMArena Vision | 1066 | — |
| Video-MME | 64.8% | — |
| GeoBench | 64% | — |
| VPCT | 34% | — |
| Blueprint-Bench 2 | — | 29.5% |
| Furniture Assembly | — | 34.2% |
Multilingual Kimi K3 leads
GPT-4o mini: 42.0 (#199), Kimi K3: 56.3 (#21)
| Benchmark | GPT-4o mini | Kimi K3 |
|---|---|---|
| LMArena Non-English | 1266 | 1466 |
| LMArena Chinese | 1265 | 1529 |
| LMArena French | 1297 | 1491 |
| LMArena German | 1272 | 1488 |
| LMArena Japanese | 1216 | 1487 |
| LMArena Korean | 1195 | 1458 |
| LMArena Russian | 1275 | 1482 |
| LMArena Spanish | 1276 | 1472 |
Instruction Following Kimi K3 leads
GPT-4o mini: 61.9 (#239), Kimi K3: 77.7 (#14)
| Benchmark | GPT-4o mini | Kimi K3 |
|---|---|---|
| LMArena Instruction Following | 1258 | 1483 |
| LiveBench Instruction Following | 56.8% | — |
| IFEval | 78.2% | — |
Long Context Kimi K3 leads
GPT-4o mini: 39.1 (#186), Kimi K3: 45.8 (#29)
| Benchmark | GPT-4o mini | Kimi K3 |
|---|---|---|
| LMArena Longer Query | 1289 | 1494 |
Writing & Preference Kimi K3 leads
GPT-4o mini: 39.5 (#248), Kimi K3: 76.6 (#4)
| Benchmark | GPT-4o mini | Kimi K3 |
|---|---|---|
| LMArena Text | 1286 | 1476 |
| LMArena Creative Writing | 1268 | 1454 |
| EQ-Bench Creative Writing | 873 | 2082 |
| LMArena Multi-Turn | 1285 | 1488 |
| Short-Story Creative Writing | 67.2% | — |
| WildBench | 79.1% | — |
| EQ-Bench 4 | — | 1339 |
| LiveBench Language | 28.6% | — |
Frequently asked questions
Is GPT-4o mini better than Kimi K3?
Kimi K3 is the stronger model overall, scoring 59.5 to 25.5 on the Noometry Index. GPT-4o mini costs 23× less per token, which makes it the better buy when Kimi K3's lead doesn't matter for your workload.
Which is cheaper, GPT-4o mini or Kimi K3?
GPT-4o mini is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; Kimi K3 lists at $3 and $15.
Is GPT-4o mini or Kimi K3 better for coding?
Kimi K3 scores higher on coding benchmarks: 61.0 versus 22.0 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-4o mini and Kimi K3 share?
30 benchmarks have published results for both models. GPT-4o mini has 60 scored results on Noometry and Kimi K3 has 53.