Model comparison
DeepSeek-R1 vs Kimi K2.5
Kimi K2.5 is the stronger model overall, scoring 48.1 to 42.3 on the Noometry Index.
Last verified . 31 shared benchmarks.
DeepSeekSummary
- They share 31 benchmarks with published results for both. DeepSeek-R1 scores higher in 0 categories and Kimi K2.5 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Kimi K2.5 leads 31.2 to 18.6.
- The biggest single-benchmark swing is ARC-AGI-1: 21.2% for DeepSeek-R1 and 65.3% for Kimi K2.5.
- Both cost about the same: $0.50 input and $2.15 output per million tokens.
- Kimi K2.5 accepts more context: 262K tokens versus 164K.
- Kimi K2.5 has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-R1 | Kimi K2.5 | |
|---|---|---|
| Provider | DeepSeek | Moonshot AI |
| Noometry Index | 42.3 | 48.1 |
| Released | 2025-01-20 | 2026-01-27 |
| Weights | Proprietary | Open |
| Context window | 164K | 262K |
| Max output | 64K | 262K |
| Input $ / M tokens | $0.50 | $0.45 |
| Output $ / M tokens | $2.15 | $2.25 |
| Results tracked | 52 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K2.5 leads
DeepSeek-R1: 46.3 (#68), Kimi K2.5: 48.8 (#53)
| Benchmark | DeepSeek-R1 | Kimi K2.5 |
|---|---|---|
| SciCode | 35.7% | 49% |
| WeirdML | 41.6% | 45.6% |
| LMArena Coding | 1427 | 1474 |
| ALE-Bench | 804.12 | 821.65 |
| SWE-bench Verified | — | 73.8% |
| SWE-bench Verified (bash only) | — | 70.8% |
| Aider Polyglot | 71.4% | — |
| LMArena WebDev | — | 1437 |
| SWE-bench Multilingual | — | 67.3% |
| LiveBench Coding | 66.7% | — |
| AlgoTune | 1.7 | — |
Agentic & Tool Use Kimi K2.5 leads
DeepSeek-R1: 30.7 (#75), Kimi K2.5: 34.2 (#48)
| Benchmark | DeepSeek-R1 | Kimi K2.5 |
|---|---|---|
| Terminal-Bench | — | 43.2% |
| DeepResearch Bench | 35.1% | — |
| OSWorld | — | 63.3% |
| BALROG | 34.9% | — |
| METR Time Horizons | 53.8% | — |
| Vending-Bench 2 | — | 1,198 |
Reasoning Kimi K2.5 leads
DeepSeek-R1: 18.6 (#278), Kimi K2.5: 31.2 (#80)
| Benchmark | DeepSeek-R1 | Kimi K2.5 |
|---|---|---|
| ARC-AGI-2 | 1.3% | 11.8% |
| SimpleBench | 40.8% | 46.8% |
| Kagi LLM Benchmark | 69.4% | 78.5% |
| ARC-AGI-1 | 21.2% | 65.3% |
| CritPt | 1.1% | 3.1% |
| LMArena Hard Prompts | 1416 | 1453 |
| Epoch Capabilities Index | 141.29 | 148.03 |
| NYT Connections (extended) | — | 69.9% |
| Chess Puzzles | — | 12% |
| EnigmaEval | — | 3.4% |
| Thematic Generalization | — | 69.4% |
| LiveBench Reasoning | 83.2% | — |
| LiveBench Data Analysis | 69.8% | — |
| ForecastBench | 60 | — |
| LiveBench | 71.6% | — |
Math Kimi K2.5 leads
DeepSeek-R1: 43.8 (#79), Kimi K2.5: 51.8 (#53)
| Benchmark | DeepSeek-R1 | Kimi K2.5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 92.2% |
| LMArena Math | 1400 | 1470 |
| MathArena Final-Answer Competitions | — | 62.3% |
| Omni-MATH | 42.4% | — |
| LiveBench Math | 80.7% | — |
| MATH Level 5 | 96.6% | — |
| FrontierMath (Feb 2025 set) | — | 27.9% |
| FrontierMath Tier 4 (v1) | — | 4.2% |
Knowledge Kimi K2.5 leads
DeepSeek-R1: 44.5 (#87), Kimi K2.5: 53.6 (#56)
| Benchmark | DeepSeek-R1 | Kimi K2.5 |
|---|---|---|
| GPQA Diamond | 76.3% | 87.6% |
| Vectara Hallucination Rate | 11.3% | 14.2% |
| LMArena Expert | 1394 | 1466 |
| Humanity's Last Exam | — | 24.4% |
| SimpleQA Verified | — | 34.3% |
| MMLU-Pro | 79.3% | — |
| Confabulations | 12.7% | — |
| GPQA (HELM) | 66.6% | — |
Multimodal Not comparable
DeepSeek-R1: —, Kimi K2.5: 41.1 (#39)
| Benchmark | DeepSeek-R1 | Kimi K2.5 |
|---|---|---|
| LMArena Vision | — | 1269 |
| LMArena Document | — | 1430 |
Multilingual Kimi K2.5 leads
DeepSeek-R1: 52.4 (#85), Kimi K2.5: 53.9 (#53)
| Benchmark | DeepSeek-R1 | Kimi K2.5 |
|---|---|---|
| LMArena Non-English | 1412 | 1433 |
| LMArena Chinese | 1442 | 1495 |
| LMArena French | 1417 | 1454 |
| LMArena German | 1404 | 1441 |
| LMArena Japanese | 1391 | 1421 |
| LMArena Korean | 1360 | 1410 |
| LMArena Russian | 1423 | 1435 |
| LMArena Spanish | 1411 | 1450 |
Instruction Following Kimi K2.5 leads
DeepSeek-R1: 72.0 (#143), Kimi K2.5: 75.3 (#64)
| Benchmark | DeepSeek-R1 | Kimi K2.5 |
|---|---|---|
| LMArena Instruction Following | 1382 | 1431 |
| LiveBench Instruction Following | 80.5% | — |
| IFEval | 78.4% | — |
Long Context Kimi K2.5 leads
DeepSeek-R1: 45.4 (#36), Kimi K2.5: 52.1 (#7)
| Benchmark | DeepSeek-R1 | Kimi K2.5 |
|---|---|---|
| Fiction.LiveBench | 75% | 86.1% |
| LMArena Longer Query | 1391 | 1445 |
| CL-bench | — | 19.3% |
| CL-bench Life | — | 13.2% |
Writing & Preference Kimi K2.5 leads
DeepSeek-R1: 61.4 (#88), Kimi K2.5: 65.1 (#53)
| Benchmark | DeepSeek-R1 | Kimi K2.5 |
|---|---|---|
| LMArena Text | 1428 | 1445 |
| LMArena Creative Writing | 1405 | 1423 |
| EQ-Bench Creative Writing | 1500 | 1579 |
| LMArena Multi-Turn | 1405 | 1444 |
| Short-Story Creative Writing | 83% | — |
| WildBench | 82.8% | — |
| LiveBench Language | 48.5% | — |
Frequently asked questions
Is DeepSeek-R1 better than Kimi K2.5?
Kimi K2.5 is the stronger model overall, scoring 48.1 to 42.3 on the Noometry Index.
Which is cheaper, DeepSeek-R1 or Kimi K2.5?
Kimi K2.5 is cheaper. It lists at $0.45 per million input tokens and $2.25 per million output tokens; DeepSeek-R1 lists at $0.50 and $2.15.
Is DeepSeek-R1 or Kimi K2.5 better for coding?
Kimi K2.5 scores higher on coding benchmarks: 48.8 versus 46.3 in the Noometry coding category.
Which has the bigger context window?
Kimi K2.5 does, with 262K tokens against 164K.
How many benchmarks do DeepSeek-R1 and Kimi K2.5 share?
31 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Kimi K2.5 has 51.