Model comparison
GPT-5.6 Sol vs Kimi K3
GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 59.5 on the Noometry Index.
Last verified . 51 shared benchmarks.
Summary
- They share 51 benchmarks with published results for both. GPT-5.6 Sol scores higher in 6 categories and Kimi K3 in 4 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5.6 Sol leads 74.8 to 63.0.
- The biggest single-benchmark swing is FrontierMath Tier 4: 82.9% for GPT-5.6 Sol and 39% for Kimi K3.
- Kimi K3 is cheaper at $3 / $15 per million input/output tokens, against $4 / $20 for GPT-5.6 Sol.
- GPT-5.6 Sol accepts more context: 1.05M tokens versus 1.05M.
- Kimi K3 has downloadable open weights; the other is API-only.
Side by side
| GPT-5.6 Sol | Kimi K3 | |
|---|---|---|
| Provider | OpenAI | Moonshot AI |
| Noometry Index | 65.0 | 59.5 |
| Released | 2026-07-09 | 2026-07-16 |
| Weights | Proprietary | Open |
| Context window | 1.05M | 1.05M |
| Max output | 128K | 1.05M |
| Input $ / M tokens | $4 | $3 |
| Output $ / M tokens | $20 | $15 |
| Results tracked | 65 | 53 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.6 Sol leads
GPT-5.6 Sol: 65.1 (#7), Kimi K3: 61.0 (#10)
| Benchmark | GPT-5.6 Sol | Kimi K3 |
|---|---|---|
| DeepSWE | 72.7% | 68.5% |
| FrontierCode | 47.5% | 44.2% |
| LMArena WebDev | 1618 | 1654 |
| FrontierSWE | 32.2% | 25.9% |
| SciCode | 57.1% | 59.5% |
| WeirdML | 89.4% | 82.6% |
| LMArena Coding | 1498 | 1508 |
| ALE-Bench | 2,177 | 1,524 |
| CursorBench | 41.7% | — |
| GSO | 76.5% | — |
| MirrorCode | 20% | — |
Agentic & Tool Use GPT-5.6 Sol leads
GPT-5.6 Sol: 50.3 (#7), Kimi K3: 41.8 (#20)
| Benchmark | GPT-5.6 Sol | Kimi K3 |
|---|---|---|
| APEX-Agents | 51.4% | 50.6% |
| τ²-bench Banking | 46.9% | 37.1% |
| PostTrainBench | 36.2% | 32% |
| GBAEval | 52.6% | 48.3% |
| GDP.pdf | 30.7% | 19% |
| Vending-Bench 2 | 9,619 | 5,165 |
| OSWorld 2.0 | 27.3% | — |
| BALROG | 60% | — |
| LMArena Search | 1257 | — |
Reasoning GPT-5.6 Sol leads
GPT-5.6 Sol: 74.8 (#8), Kimi K3: 63.0 (#17)
| Benchmark | GPT-5.6 Sol | Kimi K3 |
|---|---|---|
| ARC-AGI-2 | 92.5% | 60.4% |
| SimpleBench | 71.7% | 60.7% |
| NYT Connections (extended) | 93.8% | 93.6% |
| ARC-AGI-1 | 97.5% | 94.5% |
| CritPt | 32.3% | 23.4% |
| Chess Puzzles | 64% | 39% |
| LMArena Hard Prompts | 1484 | 1496 |
| Mystery Game Puzzles | 58% | 26% |
| DTBench | 96% | 91.2% |
| LMCA | 59.2% | 52.7% |
| Surface Evolver Bench | 93.1% | 95% |
| Epoch Capabilities Index | 161.66 | 157.45 |
| Kagi LLM Benchmark | 67% | — |
| EnigmaEval | 37.1% | — |
| EBR-Bench | 44.8% | — |
| Bench to the Future 3 | 0.14 | — |
| ForecastBench | — | 61.1 |
Math GPT-5.6 Sol leads
GPT-5.6 Sol: 85.6 (#9), Kimi K3: 74.2 (#16)
| Benchmark | GPT-5.6 Sol | Kimi K3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 89.1% | 72.2% |
| FrontierMath Tier 4 | 82.9% | 39% |
| OTIS Mock AIME 2024-2025 | 100% | 97.2% |
| ProofBench | 83% | 87% |
| LMArena Math | 1474 | 1491 |
| MathArena Final-Answer Competitions | — | 87.8% |
| FrontierMath Erdős | 0% | — |
Knowledge GPT-5.6 Sol leads
GPT-5.6 Sol: 64.3 (#18), Kimi K3: 63.2 (#21)
| Benchmark | GPT-5.6 Sol | Kimi K3 |
|---|---|---|
| GPQA Diamond | 93.5% | 93.1% |
| SimpleQA Verified | 69.7% | 50.6% |
| LMArena Expert | 1516 | 1521 |
| Vectara Hallucination Rate | 12.4% | — |
Multimodal GPT-5.6 Sol leads
GPT-5.6 Sol: 48.6 (#9), Kimi K3: 37.8 (#70)
| Benchmark | GPT-5.6 Sol | Kimi K3 |
|---|---|---|
| Blueprint-Bench 2 | 33.6% | 29.5% |
| Furniture Assembly | 56.7% | 34.2% |
| LMArena Vision | 1281 | — |
| LMArena Document | 1483 | — |
Multilingual Kimi K3 leads
GPT-5.6 Sol: 55.3 (#32), Kimi K3: 56.3 (#21)
| Benchmark | GPT-5.6 Sol | Kimi K3 |
|---|---|---|
| LMArena Non-English | 1452 | 1466 |
| LMArena Chinese | 1527 | 1529 |
| LMArena French | 1477 | 1491 |
| LMArena German | 1476 | 1488 |
| LMArena Japanese | 1471 | 1487 |
| LMArena Korean | 1442 | 1458 |
| LMArena Russian | 1468 | 1482 |
| LMArena Spanish | 1441 | 1472 |
Instruction Following Too close to call
GPT-5.6 Sol: 77.7 (#16), Kimi K3: 77.7 (#14)
| Benchmark | GPT-5.6 Sol | Kimi K3 |
|---|---|---|
| LMArena Instruction Following | 1482 | 1483 |
Long Context Too close to call
GPT-5.6 Sol: 45.4 (#42), Kimi K3: 45.8 (#29)
| Benchmark | GPT-5.6 Sol | Kimi K3 |
|---|---|---|
| LMArena Longer Query | 1480 | 1494 |
Writing & Preference Kimi K3 leads
GPT-5.6 Sol: 73.3 (#12), Kimi K3: 76.6 (#4)
| Benchmark | GPT-5.6 Sol | Kimi K3 |
|---|---|---|
| LMArena Text | 1457 | 1476 |
| LMArena Creative Writing | 1448 | 1454 |
| EQ-Bench Creative Writing | 1972 | 2082 |
| EQ-Bench 4 | 1250 | 1339 |
| LMArena Multi-Turn | 1460 | 1488 |
Frequently asked questions
Is GPT-5.6 Sol better than Kimi K3?
GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 59.5 on the Noometry Index.
Which is cheaper, GPT-5.6 Sol or Kimi K3?
Kimi K3 is cheaper. It lists at $3 per million input tokens and $15 per million output tokens; GPT-5.6 Sol lists at $4 and $20.
Is GPT-5.6 Sol or Kimi K3 better for coding?
GPT-5.6 Sol scores higher on coding benchmarks: 65.1 versus 61.0 in the Noometry coding category.
Which has the bigger context window?
GPT-5.6 Sol does, with 1.05M tokens against 1.05M.
How many benchmarks do GPT-5.6 Sol and Kimi K3 share?
51 benchmarks have published results for both models. GPT-5.6 Sol has 65 scored results on Noometry and Kimi K3 has 53.