Model comparison
DeepSeek-V3.1 vs Kimi K3
Kimi K3 is the stronger model overall, scoring 59.5 to 42.8 on the Noometry Index. DeepSeek-V3.1 costs 14× less per token, which makes it the better buy when Kimi K3's lead doesn't matter for your workload.
Last verified . 24 shared benchmarks.
DeepSeekSummary
- They share 24 benchmarks with published results for both. DeepSeek-V3.1 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 38.9.
- The biggest single-benchmark swing is WeirdML: 38.4% for DeepSeek-V3.1 and 82.6% for Kimi K3.
- DeepSeek-V3.1 is cheaper at $0.25 / $0.95 per million input/output tokens, against $3 / $15 for Kimi K3.
- Kimi K3 accepts more context: 1.05M tokens versus 164K.
Side by side
| DeepSeek-V3.1 | Kimi K3 | |
|---|---|---|
| Provider | DeepSeek | Moonshot AI |
| Noometry Index | 42.8 | 59.5 |
| Released | 2025-08-21 | 2026-07-16 |
| Weights | Open | Open |
| Context window | 164K | 1.05M |
| Max output | 8K | 1.05M |
| Input $ / M tokens | $0.25 | $3 |
| Output $ / M tokens | $0.95 | $15 |
| Results tracked | 27 | 53 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K3 leads
DeepSeek-V3.1: 40.3 (#144), Kimi K3: 61.0 (#10)
| Benchmark | DeepSeek-V3.1 | Kimi K3 |
|---|---|---|
| WeirdML | 38.4% | 82.6% |
| LMArena Coding | 1417 | 1508 |
| DeepSWE | — | 68.5% |
| FrontierCode | — | 44.2% |
| LMArena WebDev | — | 1654 |
| FrontierSWE | — | 25.9% |
| SciCode | — | 59.5% |
| ALE-Bench | — | 1,524 |
Agentic & Tool Use Not comparable
DeepSeek-V3.1: —, Kimi K3: 41.8 (#20)
| Benchmark | DeepSeek-V3.1 | Kimi K3 |
|---|---|---|
| APEX-Agents | — | 50.6% |
| τ²-bench Banking | — | 37.1% |
| PostTrainBench | — | 32% |
| GBAEval | — | 48.3% |
| GDP.pdf | — | 19% |
| Vending-Bench 2 | — | 5,165 |
Reasoning Kimi K3 leads
DeepSeek-V3.1: 27.9 (#110), Kimi K3: 63.0 (#17)
| Benchmark | DeepSeek-V3.1 | Kimi K3 |
|---|---|---|
| SimpleBench | 40% | 60.7% |
| LMArena Hard Prompts | 1417 | 1496 |
| DTBench | 82.7% | 91.2% |
| LMCA | 24.3% | 52.7% |
| Epoch Capabilities Index | 139.92 | 157.45 |
| ForecastBench | 58 | 61.1 |
| ARC-AGI-2 | — | 60.4% |
| Kagi LLM Benchmark | 53.2% | — |
| NYT Connections (extended) | — | 93.6% |
| ARC-AGI-1 | — | 94.5% |
| CritPt | — | 23.4% |
| Chess Puzzles | — | 39% |
| Mystery Game Puzzles | — | 26% |
| Surface Evolver Bench | — | 95% |
Math Kimi K3 leads
DeepSeek-V3.1: 38.9 (#122), Kimi K3: 74.2 (#16)
| Benchmark | DeepSeek-V3.1 | Kimi K3 |
|---|---|---|
| LMArena Math | 1420 | 1491 |
| FrontierMath (Tiers 1-3) | — | 72.2% |
| FrontierMath Tier 4 | — | 39% |
| MathArena Final-Answer Competitions | — | 87.8% |
| OTIS Mock AIME 2024-2025 | — | 97.2% |
| ProofBench | — | 87% |
Knowledge Kimi K3 leads
DeepSeek-V3.1: 43.7 (#90), Kimi K3: 63.2 (#21)
| Benchmark | DeepSeek-V3.1 | Kimi K3 |
|---|---|---|
| LMArena Expert | 1405 | 1521 |
| GPQA Diamond | — | 93.1% |
| SimpleQA Verified | — | 50.6% |
| Vectara Hallucination Rate | 5.5% | — |
Multimodal Not comparable
DeepSeek-V3.1: —, Kimi K3: 37.8 (#70)
| Benchmark | DeepSeek-V3.1 | Kimi K3 |
|---|---|---|
| Blueprint-Bench 2 | — | 29.5% |
| Furniture Assembly | — | 34.2% |
Multilingual Kimi K3 leads
DeepSeek-V3.1: 51.6 (#106), Kimi K3: 56.3 (#21)
| Benchmark | DeepSeek-V3.1 | Kimi K3 |
|---|---|---|
| LMArena Non-English | 1400 | 1466 |
| LMArena Chinese | 1469 | 1529 |
| LMArena French | 1447 | 1491 |
| LMArena German | 1411 | 1488 |
| LMArena Japanese | 1378 | 1487 |
| LMArena Korean | 1337 | 1458 |
| LMArena Russian | 1405 | 1482 |
| LMArena Spanish | 1431 | 1472 |
Instruction Following Kimi K3 leads
DeepSeek-V3.1: 73.9 (#110), Kimi K3: 77.7 (#14)
| Benchmark | DeepSeek-V3.1 | Kimi K3 |
|---|---|---|
| LMArena Instruction Following | 1400 | 1483 |
Long Context Kimi K3 leads
DeepSeek-V3.1: 36.3 (#232), Kimi K3: 45.8 (#29)
| Benchmark | DeepSeek-V3.1 | Kimi K3 |
|---|---|---|
| LMArena Longer Query | 1422 | 1494 |
| Fiction.LiveBench | 52.8% | — |
Writing & Preference Kimi K3 leads
DeepSeek-V3.1: 60.3 (#98), Kimi K3: 76.6 (#4)
| Benchmark | DeepSeek-V3.1 | Kimi K3 |
|---|---|---|
| LMArena Text | 1420 | 1476 |
| LMArena Creative Writing | 1401 | 1454 |
| EQ-Bench Creative Writing | 1436 | 2082 |
| LMArena Multi-Turn | 1408 | 1488 |
| EQ-Bench 4 | — | 1339 |
Frequently asked questions
Is DeepSeek-V3.1 better than Kimi K3?
Kimi K3 is the stronger model overall, scoring 59.5 to 42.8 on the Noometry Index. DeepSeek-V3.1 costs 14× less per token, which makes it the better buy when Kimi K3's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-V3.1 or Kimi K3?
DeepSeek-V3.1 is cheaper. It lists at $0.25 per million input tokens and $0.95 per million output tokens; Kimi K3 lists at $3 and $15.
Is DeepSeek-V3.1 or Kimi K3 better for coding?
Kimi K3 scores higher on coding benchmarks: 61.0 versus 40.3 in the Noometry coding category.
Which has the bigger context window?
Kimi K3 does, with 1.05M tokens against 164K.
How many benchmarks do DeepSeek-V3.1 and Kimi K3 share?
24 benchmarks have published results for both models. DeepSeek-V3.1 has 27 scored results on Noometry and Kimi K3 has 53.