Model comparison
DeepSeek-V2.5 (Sep 2024) vs Kimi K3
Kimi K3 is the stronger model overall, scoring 59.5 to 37.6 on the Noometry Index.
Last verified . 17 shared benchmarks.
DeepSeekSummary
- They share 17 benchmarks with published results for both. DeepSeek-V2.5 (Sep 2024) 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 35.9.
Side by side
| DeepSeek-V2.5 (Sep 2024) | Kimi K3 | |
|---|---|---|
| Provider | DeepSeek | Moonshot AI |
| Noometry Index | 37.6 | 59.5 |
| Released | 2024-09-06 | 2026-07-16 |
| Weights | Open | Open |
| Context window | — | 1.05M |
| Max output | — | 1.05M |
| Input $ / M tokens | — | $3 |
| Output $ / M tokens | — | $15 |
| Results tracked | 22 | 53 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K3 leads
DeepSeek-V2.5 (Sep 2024): 31.7 (#281), Kimi K3: 61.0 (#10)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Kimi K3 |
|---|---|---|
| LMArena Coding | 1309 | 1508 |
| DeepSWE | — | 68.5% |
| FrontierCode | — | 44.2% |
| Aider Polyglot | 17.8% | — |
| LMArena WebDev | — | 1654 |
| FrontierSWE | — | 25.9% |
| SciCode | — | 59.5% |
| WeirdML | — | 82.6% |
| BigCodeBench Instruct | 48.6% | — |
| BigCodeBench Complete | 53.2% | — |
| ALE-Bench | — | 1,524 |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |
Agentic & Tool Use Not comparable
DeepSeek-V2.5 (Sep 2024): —, Kimi K3: 41.8 (#20)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | 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-V2.5 (Sep 2024): 25.6 (#145), Kimi K3: 63.0 (#17)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Kimi K3 |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1496 |
| ARC-AGI-2 | — | 60.4% |
| SimpleBench | — | 60.7% |
| NYT Connections (extended) | — | 93.6% |
| ARC-AGI-1 | — | 94.5% |
| CritPt | — | 23.4% |
| Chess Puzzles | — | 39% |
| Mystery Game Puzzles | — | 26% |
| DTBench | — | 91.2% |
| LMCA | — | 52.7% |
| Surface Evolver Bench | — | 95% |
| Epoch Capabilities Index | — | 157.45 |
| ForecastBench | — | 61.1 |
Math Kimi K3 leads
DeepSeek-V2.5 (Sep 2024): 35.9 (#177), Kimi K3: 74.2 (#16)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Kimi K3 |
|---|---|---|
| LMArena Math | 1288 | 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-V2.5 (Sep 2024): 34.8 (#193), Kimi K3: 63.2 (#21)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Kimi K3 |
|---|---|---|
| LMArena Expert | 1266 | 1521 |
| GPQA Diamond | — | 93.1% |
| SimpleQA Verified | — | 50.6% |
Multimodal Not comparable
DeepSeek-V2.5 (Sep 2024): —, Kimi K3: 37.8 (#70)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Kimi K3 |
|---|---|---|
| Blueprint-Bench 2 | — | 29.5% |
| Furniture Assembly | — | 34.2% |
Multilingual Kimi K3 leads
DeepSeek-V2.5 (Sep 2024): 42.5 (#193), Kimi K3: 56.3 (#21)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Kimi K3 |
|---|---|---|
| LMArena Non-English | 1273 | 1466 |
| LMArena Chinese | 1318 | 1529 |
| LMArena French | 1289 | 1491 |
| LMArena German | 1258 | 1488 |
| LMArena Japanese | 1228 | 1487 |
| LMArena Korean | 1209 | 1458 |
| LMArena Russian | 1289 | 1482 |
| LMArena Spanish | 1248 | 1472 |
Instruction Following Kimi K3 leads
DeepSeek-V2.5 (Sep 2024): 67.5 (#194), Kimi K3: 77.7 (#14)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Kimi K3 |
|---|---|---|
| LMArena Instruction Following | 1280 | 1483 |
Long Context Kimi K3 leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#174), Kimi K3: 45.8 (#29)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Kimi K3 |
|---|---|---|
| LMArena Longer Query | 1301 | 1494 |
Writing & Preference Kimi K3 leads
DeepSeek-V2.5 (Sep 2024): 49.8 (#187), Kimi K3: 76.6 (#4)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Kimi K3 |
|---|---|---|
| LMArena Text | 1294 | 1476 |
| LMArena Creative Writing | 1285 | 1454 |
| LMArena Multi-Turn | 1297 | 1488 |
| EQ-Bench Creative Writing | — | 2082 |
| EQ-Bench 4 | — | 1339 |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than Kimi K3?
Kimi K3 is the stronger model overall, scoring 59.5 to 37.6 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or Kimi K3 better for coding?
Kimi K3 scores higher on coding benchmarks: 61.0 versus 31.7 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and Kimi K3 share?
17 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and Kimi K3 has 53.