Model comparison
GPT-5.4 vs Kimi K3
GPT-5.4 and Kimi K3 score almost the same on the Noometry Index (59.4 vs 59.5), so choose on price, context window or the category you care about most.
Last verified . 48 shared benchmarks.
Summary
- They share 48 benchmarks with published results for both. GPT-5.4 scores higher in 4 categories and Kimi K3 in 6 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in coding, where Kimi K3 leads 61.0 to 52.6.
- The biggest single-benchmark swing is ProofBench: 56% for GPT-5.4 and 87% for Kimi K3.
- GPT-5.4 is cheaper at $2.50 / $15 per million input/output tokens, against $3 / $15 for Kimi K3.
- GPT-5.4 accepts more context: 1.05M tokens versus 1.05M.
- Kimi K3 has downloadable open weights; the other is API-only.
Side by side
| GPT-5.4 | Kimi K3 | |
|---|---|---|
| Provider | OpenAI | Moonshot AI |
| Noometry Index | 59.4 | 59.5 |
| Released | 2026-03-05 | 2026-07-16 |
| Weights | Proprietary | Open |
| Context window | 1.05M | 1.05M |
| Max output | 128K | 1.05M |
| Input $ / M tokens | $2.50 | $3 |
| Output $ / M tokens | $15 | $15 |
| Results tracked | 68 | 53 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K3 leads
GPT-5.4: 52.6 (#33), Kimi K3: 61.0 (#10)
| Benchmark | GPT-5.4 | Kimi K3 |
|---|---|---|
| DeepSWE | 51.8% | 68.5% |
| LMArena WebDev | 1465 | 1654 |
| SciCode | 56.6% | 59.5% |
| WeirdML | 77.7% | 82.6% |
| LMArena Coding | 1497 | 1508 |
| ALE-Bench | 1,607 | 1,524 |
| SWE-bench Verified | 76.9% | — |
| FrontierCode | — | 44.2% |
| FrontierSWE | — | 25.9% |
| GSO | 31.4% | — |
| MirrorCode | 15.6% | — |
| AlgoTune | 1.85 | — |
Agentic & Tool Use GPT-5.4 leads
GPT-5.4: 46.5 (#13), Kimi K3: 41.8 (#20)
| Benchmark | GPT-5.4 | Kimi K3 |
|---|---|---|
| APEX-Agents | 52.4% | 50.6% |
| τ²-bench Banking | 39.4% | 37.1% |
| PostTrainBench | 19% | 32% |
| GBAEval | 45.1% | 48.3% |
| Vending-Bench 2 | 6,144 | 5,165 |
| Terminal-Bench | 81.8% | — |
| DeepResearch Bench | 35.1% | — |
| GDP.pdf | — | 19% |
| LMArena Search | 1197 | — |
| METR Time Horizons | 74.3% | — |
Reasoning Kimi K3 leads
GPT-5.4: 61.8 (#19), Kimi K3: 63.0 (#17)
| Benchmark | GPT-5.4 | Kimi K3 |
|---|---|---|
| ARC-AGI-2 | 74% | 60.4% |
| NYT Connections (extended) | 91.3% | 93.6% |
| ARC-AGI-1 | 93.7% | 94.5% |
| CritPt | 23.4% | 23.4% |
| Chess Puzzles | 44% | 39% |
| LMArena Hard Prompts | 1485 | 1496 |
| Mystery Game Puzzles | 37% | 26% |
| DTBench | 94.4% | 91.2% |
| LMCA | 52% | 52.7% |
| Epoch Capabilities Index | 156.81 | 157.45 |
| ForecastBench | 59.5 | 61.1 |
| SimpleBench | — | 60.7% |
| Kagi LLM Benchmark | 63.8% | — |
| EnigmaEval | 16% | — |
| Thematic Generalization | 80% | — |
| EBR-Bench | 25.4% | — |
| Surface Evolver Bench | — | 95% |
Math Too close to call
GPT-5.4: 73.5 (#19), Kimi K3: 74.2 (#16)
| Benchmark | GPT-5.4 | Kimi K3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 78.6% | 72.2% |
| FrontierMath Tier 4 | 49% | 39% |
| MathArena Final-Answer Competitions | 83.1% | 87.8% |
| OTIS Mock AIME 2024-2025 | 97.8% | 97.2% |
| ProofBench | 56% | 87% |
| LMArena Math | 1488 | 1491 |
| FrontierMath (Feb 2025 set) | 47.6% | — |
| FrontierMath Tier 4 (v1) | 27.1% | — |
Knowledge GPT-5.4 leads
GPT-5.4: 65.3 (#14), Kimi K3: 63.2 (#21)
| Benchmark | GPT-5.4 | Kimi K3 |
|---|---|---|
| GPQA Diamond | 93.3% | 93.1% |
| SimpleQA Verified | 45.1% | 50.6% |
| LMArena Expert | 1507 | 1521 |
| Humanity's Last Exam | 36.2% | — |
| Vectara Hallucination Rate | 7% | — |
Multimodal GPT-5.4 leads
GPT-5.4: 43.7 (#20), Kimi K3: 37.8 (#70)
| Benchmark | GPT-5.4 | Kimi K3 |
|---|---|---|
| Blueprint-Bench 2 | 27.1% | 29.5% |
| Furniture Assembly | 37.5% | 34.2% |
| LMArena Vision | 1303 | — |
| LMArena Document | 1471 | — |
Multilingual Too close to call
GPT-5.4: 56.2 (#23), Kimi K3: 56.3 (#21)
| Benchmark | GPT-5.4 | Kimi K3 |
|---|---|---|
| LMArena Non-English | 1465 | 1466 |
| LMArena Chinese | 1519 | 1529 |
| LMArena French | 1493 | 1491 |
| LMArena German | 1472 | 1488 |
| LMArena Japanese | 1485 | 1487 |
| LMArena Korean | 1448 | 1458 |
| LMArena Russian | 1480 | 1482 |
| LMArena Spanish | 1454 | 1472 |
Instruction Following Too close to call
GPT-5.4: 77.1 (#27), Kimi K3: 77.7 (#14)
| Benchmark | GPT-5.4 | Kimi K3 |
|---|---|---|
| LMArena Instruction Following | 1469 | 1483 |
Long Context GPT-5.4 leads
GPT-5.4: 50.3 (#8), Kimi K3: 45.8 (#29)
| Benchmark | GPT-5.4 | Kimi K3 |
|---|---|---|
| LMArena Longer Query | 1473 | 1494 |
| CL-bench | 27.9% | — |
| CL-bench Life | 21.7% | — |
Writing & Preference Kimi K3 leads
GPT-5.4: 71.9 (#17), Kimi K3: 76.6 (#4)
| Benchmark | GPT-5.4 | Kimi K3 |
|---|---|---|
| LMArena Text | 1469 | 1476 |
| LMArena Creative Writing | 1439 | 1454 |
| EQ-Bench Creative Writing | 1840 | 2082 |
| EQ-Bench 4 | 1272 | 1339 |
| LMArena Multi-Turn | 1482 | 1488 |
Frequently asked questions
Is GPT-5.4 better than Kimi K3?
GPT-5.4 and Kimi K3 score almost the same on the Noometry Index (59.4 vs 59.5), so choose on price, context window or the category you care about most.
Which is cheaper, GPT-5.4 or Kimi K3?
GPT-5.4 is cheaper. It lists at $2.50 per million input tokens and $15 per million output tokens; Kimi K3 lists at $3 and $15.
Is GPT-5.4 or Kimi K3 better for coding?
Kimi K3 scores higher on coding benchmarks: 61.0 versus 52.6 in the Noometry coding category.
Which has the bigger context window?
GPT-5.4 does, with 1.05M tokens against 1.05M.
How many benchmarks do GPT-5.4 and Kimi K3 share?
48 benchmarks have published results for both models. GPT-5.4 has 68 scored results on Noometry and Kimi K3 has 53.