Model comparison
DeepSeek-V3.1-Terminus vs Kimi K2 (Jul 2025)
DeepSeek-V3.1-Terminus is the stronger model overall, scoring 42.6 to 40.9 on the Noometry Index.
Last verified . 12 shared benchmarks.
DeepSeekSummary
- They share 12 benchmarks with published results for both. DeepSeek-V3.1-Terminus scores higher in 4 categories and Kimi K2 (Jul 2025) in 3 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K2 (Jul 2025) leads 40.1 to 35.7.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 57.4% for DeepSeek-V3.1-Terminus and 64.4% for Kimi K2 (Jul 2025).
- DeepSeek-V3.1-Terminus is cheaper at $0.30 / $1 per million input/output tokens, against $0.57 / $2.30 for Kimi K2 (Jul 2025).
- Kimi K2 (Jul 2025) accepts more context: 262K tokens versus 164K.
Side by side
| DeepSeek-V3.1-Terminus | Kimi K2 (Jul 2025) | |
|---|---|---|
| Provider | DeepSeek | Moonshot AI |
| Noometry Index | 42.6 | 40.9 |
| Released | 2025-09-22 | 2025-07-12 |
| Weights | Open | Open |
| Context window | 164K | 262K |
| Max output | 66K | 262K |
| Input $ / M tokens | $0.30 | $0.57 |
| Output $ / M tokens | $1 | $2.30 |
| Results tracked | 16 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K2 (Jul 2025) leads
DeepSeek-V3.1-Terminus: 41.5 (#114), Kimi K2 (Jul 2025): 43.1 (#83)
| Benchmark | DeepSeek-V3.1-Terminus | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Coding | 1426 | 1399 |
| ALE-Bench | 745.17 | 597.5 |
| SWE-bench Verified (bash only) | — | 63.4% |
| Aider Polyglot | — | 59.1% |
| SciCode | 40.6% | — |
| GSO | — | 4.9% |
| WeirdML | — | 42.8% |
Agentic & Tool Use Not comparable
DeepSeek-V3.1-Terminus: —, Kimi K2 (Jul 2025): 32.2 (#59)
| Benchmark | DeepSeek-V3.1-Terminus | Kimi K2 (Jul 2025) |
|---|---|---|
| Terminal-Bench | — | 35.7% |
| Berkeley Function Calling Leaderboard | — | 59.1% |
| METR Time Horizons | — | 59.2% |
Reasoning DeepSeek-V3.1-Terminus leads
DeepSeek-V3.1-Terminus: 25.7 (#126), Kimi K2 (Jul 2025): 22.7 (#173)
| Benchmark | DeepSeek-V3.1-Terminus | Kimi K2 (Jul 2025) |
|---|---|---|
| Kagi LLM Benchmark | 57.4% | 64.4% |
| LMArena Hard Prompts | 1427 | 1384 |
| SimpleBench | — | 26.3% |
| CritPt | 1.7% | — |
| DTBench | 81.3% | — |
| LMCA | 28.6% | — |
| Epoch Capabilities Index | — | 146.02 |
| ForecastBench | — | 60.2 |
Math Kimi K2 (Jul 2025) leads
DeepSeek-V3.1-Terminus: 35.7 (#138), Kimi K2 (Jul 2025): 40.1 (#81)
| Benchmark | DeepSeek-V3.1-Terminus | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Math | 1402 | 1396 |
| Omni-MATH | — | 65.4% |
| FrontierMath (Feb 2025 set) | — | 21.4% |
| FrontierMath Tier 4 (v1) | — | 0% |
Knowledge Not comparable
DeepSeek-V3.1-Terminus: —, Kimi K2 (Jul 2025): 36.9 (#153)
| Benchmark | DeepSeek-V3.1-Terminus | Kimi K2 (Jul 2025) |
|---|---|---|
| MMLU-Pro | — | 81.9% |
| Confabulations | — | 20.4% |
| Vectara Hallucination Rate | — | 17.9% |
| GPQA (HELM) | — | 65.3% |
| LMArena Expert | — | 1363 |
Multilingual DeepSeek-V3.1-Terminus leads
DeepSeek-V3.1-Terminus: 52.3 (#87), Kimi K2 (Jul 2025): 49.8 (#125)
| Benchmark | DeepSeek-V3.1-Terminus | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Non-English | 1407 | 1372 |
| LMArena Russian | 1434 | 1384 |
| LMArena Chinese | — | 1416 |
| LMArena French | — | 1379 |
| LMArena German | — | 1386 |
| LMArena Japanese | — | 1349 |
| LMArena Korean | — | 1324 |
| LMArena Spanish | — | 1383 |
Instruction Following DeepSeek-V3.1-Terminus leads
DeepSeek-V3.1-Terminus: 74.4 (#96), Kimi K2 (Jul 2025): 71.3 (#151)
| Benchmark | DeepSeek-V3.1-Terminus | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Instruction Following | 1406 | 1347 |
| IFEval | — | 85% |
Long Context DeepSeek-V3.1-Terminus leads
DeepSeek-V3.1-Terminus: 43.5 (#88), Kimi K2 (Jul 2025): 41.2 (#140)
| Benchmark | DeepSeek-V3.1-Terminus | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Longer Query | 1426 | 1353 |
| Fiction.LiveBench | — | 66.7% |
| CL-bench | — | 17.6% |
Writing & Preference Kimi K2 (Jul 2025) leads
DeepSeek-V3.1-Terminus: 61.3 (#87), Kimi K2 (Jul 2025): 62.5 (#76)
| Benchmark | DeepSeek-V3.1-Terminus | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Text | 1420 | 1380 |
| LMArena Creative Writing | 1404 | 1349 |
| LMArena Multi-Turn | 1411 | 1371 |
| Short-Story Creative Writing | — | 85.6% |
| EQ-Bench Creative Writing | — | 1666 |
| WildBench | — | 86.2% |
Frequently asked questions
Is DeepSeek-V3.1-Terminus better than Kimi K2 (Jul 2025)?
DeepSeek-V3.1-Terminus is the stronger model overall, scoring 42.6 to 40.9 on the Noometry Index.
Which is cheaper, DeepSeek-V3.1-Terminus or Kimi K2 (Jul 2025)?
DeepSeek-V3.1-Terminus is cheaper. It lists at $0.30 per million input tokens and $1 per million output tokens; Kimi K2 (Jul 2025) lists at $0.57 and $2.30.
Is DeepSeek-V3.1-Terminus or Kimi K2 (Jul 2025) better for coding?
Kimi K2 (Jul 2025) scores higher on coding benchmarks: 43.1 versus 41.5 in the Noometry coding category.
Which has the bigger context window?
Kimi K2 (Jul 2025) does, with 262K tokens against 164K.
How many benchmarks do DeepSeek-V3.1-Terminus and Kimi K2 (Jul 2025) share?
12 benchmarks have published results for both models. DeepSeek-V3.1-Terminus has 16 scored results on Noometry and Kimi K2 (Jul 2025) has 42.