Model comparison
Grok 4 vs Kimi K2.5
Grok 4 and Kimi K2.5 score almost the same on the Noometry Index (48.1 vs 48.1), so choose on price, context window or the category you care about most.
Last verified . 31 shared benchmarks.
Summary
- They share 31 benchmarks with published results for both. Grok 4 scores higher in 5 categories and Kimi K2.5 in 5 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in long context, where Grok 4 leads 63.1 to 52.1.
- The biggest single-benchmark swing is Terminal-Bench: 27.2% for Grok 4 and 43.2% for Kimi K2.5.
- Kimi K2.5 has downloadable open weights; the other is API-only.
Side by side
| Grok 4 | Kimi K2.5 | |
|---|---|---|
| Provider | xAI | Moonshot AI |
| Noometry Index | 48.1 | 48.1 |
| Released | 2025-07-09 | 2026-01-27 |
| Weights | Proprietary | Open |
| Context window | — | 262K |
| Max output | — | 262K |
| Input $ / M tokens | — | $0.45 |
| Output $ / M tokens | — | $2.25 |
| Results tracked | 48 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok 4 leads
Grok 4: 50.3 (#46), Kimi K2.5: 48.8 (#53)
| Benchmark | Grok 4 | Kimi K2.5 |
|---|---|---|
| WeirdML | 45.7% | 45.6% |
| LMArena Coding | 1408 | 1474 |
| SWE-bench Verified | — | 73.8% |
| SWE-bench Verified (bash only) | — | 70.8% |
| Aider Polyglot | 79.6% | — |
| LMArena WebDev | — | 1437 |
| SWE-bench Multilingual | — | 67.3% |
| SciCode | — | 49% |
| ALE-Bench | — | 821.65 |
Agentic & Tool Use Kimi K2.5 leads
Grok 4: 32.3 (#68), Kimi K2.5: 34.2 (#48)
| Benchmark | Grok 4 | Kimi K2.5 |
|---|---|---|
| Terminal-Bench | 27.2% | 43.2% |
| Berkeley Function Calling Leaderboard | 63% | — |
| GDPval | 21.1% | — |
| Cybench | 43% | — |
| DeepResearch Bench | 47.3% | — |
| OSWorld | — | 63.3% |
| BALROG | 43.6% | — |
| LMArena Search | 1142 | — |
| METR Time Horizons | 66.6% | — |
| Vending-Bench 2 | — | 1,198 |
Reasoning Grok 4 leads
Grok 4: 36.7 (#65), Kimi K2.5: 31.2 (#80)
| Benchmark | Grok 4 | Kimi K2.5 |
|---|---|---|
| ARC-AGI-2 | 16% | 11.8% |
| SimpleBench | 60.5% | 46.8% |
| Kagi LLM Benchmark | 73.6% | 78.5% |
| ARC-AGI-1 | 66.7% | 65.3% |
| Chess Puzzles | 28% | 12% |
| LMArena Hard Prompts | 1409 | 1453 |
| Epoch Capabilities Index | 146.44 | 148.03 |
| NYT Connections (extended) | — | 69.9% |
| CritPt | — | 3.1% |
| EnigmaEval | — | 3.4% |
| Thematic Generalization | — | 69.4% |
| ForecastBench | 60.9 | — |
Math Kimi K2.5 leads
Grok 4: 48.4 (#64), Kimi K2.5: 51.8 (#53)
| Benchmark | Grok 4 | Kimi K2.5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 84% | 92.2% |
| LMArena Math | 1422 | 1470 |
| FrontierMath (Feb 2025 set) | 19.7% | 27.9% |
| FrontierMath Tier 4 (v1) | 2.1% | 4.2% |
| MathArena Final-Answer Competitions | — | 62.3% |
| Omni-MATH | 60.3% | — |
Knowledge Too close to call
Grok 4: 53.8 (#55), Kimi K2.5: 53.6 (#56)
| Benchmark | Grok 4 | Kimi K2.5 |
|---|---|---|
| GPQA Diamond | 87% | 87.6% |
| LMArena Expert | 1415 | 1466 |
| Humanity's Last Exam | — | 24.4% |
| SimpleQA Verified | — | 34.3% |
| MMLU-Pro | 85.1% | — |
| Confabulations | 12.4% | — |
| Vectara Hallucination Rate | — | 14.2% |
| GPQA (HELM) | 72.7% | — |
Multimodal Kimi K2.5 leads
Grok 4: 33.7 (#94), Kimi K2.5: 41.1 (#39)
| Benchmark | Grok 4 | Kimi K2.5 |
|---|---|---|
| LMArena Vision | 1210 | 1269 |
| GeoBench | 45% | — |
| LMArena Document | — | 1430 |
Multilingual Kimi K2.5 leads
Grok 4: 51.8 (#103), Kimi K2.5: 53.9 (#53)
| Benchmark | Grok 4 | Kimi K2.5 |
|---|---|---|
| LMArena Non-English | 1403 | 1433 |
| LMArena Chinese | 1427 | 1495 |
| LMArena French | 1418 | 1454 |
| LMArena German | 1429 | 1441 |
| LMArena Japanese | 1394 | 1421 |
| LMArena Korean | 1377 | 1410 |
| LMArena Russian | 1410 | 1435 |
| LMArena Spanish | 1420 | 1450 |
Instruction Following Grok 4 leads
Grok 4: 79.2 (#5), Kimi K2.5: 75.3 (#64)
| Benchmark | Grok 4 | Kimi K2.5 |
|---|---|---|
| LMArena Instruction Following | 1387 | 1431 |
| IFEval | 94.9% | — |
Long Context Grok 4 leads
Grok 4: 63.1 (#4), Kimi K2.5: 52.1 (#7)
| Benchmark | Grok 4 | Kimi K2.5 |
|---|---|---|
| Fiction.LiveBench | 94.4% | 86.1% |
| LMArena Longer Query | 1409 | 1445 |
| CL-bench | — | 19.3% |
| CL-bench Life | — | 13.2% |
Writing & Preference Kimi K2.5 leads
Grok 4: 58.5 (#116), Kimi K2.5: 65.1 (#53)
| Benchmark | Grok 4 | Kimi K2.5 |
|---|---|---|
| LMArena Text | 1411 | 1445 |
| LMArena Creative Writing | 1397 | 1423 |
| LMArena Multi-Turn | 1416 | 1444 |
| Short-Story Creative Writing | 76.9% | — |
| EQ-Bench Creative Writing | — | 1579 |
| WildBench | 79.7% | — |
Frequently asked questions
Is Grok 4 better than Kimi K2.5?
Grok 4 and Kimi K2.5 score almost the same on the Noometry Index (48.1 vs 48.1), so choose on price, context window or the category you care about most.
Is Grok 4 or Kimi K2.5 better for coding?
Grok 4 scores higher on coding benchmarks: 50.3 versus 48.8 in the Noometry coding category.
How many benchmarks do Grok 4 and Kimi K2.5 share?
31 benchmarks have published results for both models. Grok 4 has 48 scored results on Noometry and Kimi K2.5 has 51.