Model comparison
Grok 4.6 vs Kimi K2.5
Grok 4.6 is the stronger model overall, scoring 56.9 to 48.1 on the Noometry Index. Kimi K2.5 costs 3.3× less per token, which makes it the better buy when Grok 4.6's lead doesn't matter for your workload.
Last verified . 34 shared benchmarks.
Summary
- They share 34 benchmarks with published results for both. Grok 4.6 scores higher in 7 categories and Kimi K2.5 in 3 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4.6 leads 61.4 to 31.2.
- The biggest single-benchmark swing is ARC-AGI-2: 67.1% for Grok 4.6 and 11.8% for Kimi K2.5.
- Kimi K2.5 is cheaper at $0.45 / $2.25 per million input/output tokens, against $2 / $6 for Grok 4.6.
- Grok 4.6 accepts more context: 500K tokens versus 262K.
- Kimi K2.5 has downloadable open weights; the other is API-only.
Side by side
| Grok 4.6 | Kimi K2.5 | |
|---|---|---|
| Provider | xAI | Moonshot AI |
| Noometry Index | 56.9 | 48.1 |
| Released | 2026-08-12 | 2026-01-27 |
| Weights | Proprietary | Open |
| Context window | 500K | 262K |
| Max output | 500K | 262K |
| Input $ / M tokens | $2 | $0.45 |
| Output $ / M tokens | $6 | $2.25 |
| Results tracked | 49 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok 4.6 leads
Grok 4.6: 58.5 (#16), Kimi K2.5: 48.8 (#53)
| Benchmark | Grok 4.6 | Kimi K2.5 |
|---|---|---|
| LMArena WebDev | 1617 | 1437 |
| SciCode | 56.5% | 49% |
| WeirdML | 67.3% | 45.6% |
| LMArena Coding | 1465 | 1474 |
| ALE-Bench | 1,508 | 821.65 |
| SWE-bench Verified | — | 73.8% |
| DeepSWE | 67.5% | — |
| FrontierCode | 48% | — |
| SWE-bench Verified (bash only) | — | 70.8% |
| CursorBench | 41.4% | — |
| SWE-bench Multilingual | — | 67.3% |
| FrontierSWE | 25.3% | — |
Agentic & Tool Use Grok 4.6 leads
Grok 4.6: 39.4 (#27), Kimi K2.5: 34.2 (#48)
| Benchmark | Grok 4.6 | Kimi K2.5 |
|---|---|---|
| Vending-Bench 2 | 9,047 | 1,198 |
| Terminal-Bench | — | 43.2% |
| APEX-Agents | 65.3% | — |
| OSWorld | — | 63.3% |
| GDP.pdf | 17.2% | — |
Reasoning Grok 4.6 leads
Grok 4.6: 61.4 (#20), Kimi K2.5: 31.2 (#80)
| Benchmark | Grok 4.6 | Kimi K2.5 |
|---|---|---|
| ARC-AGI-2 | 67.1% | 11.8% |
| SimpleBench | 75.9% | 46.8% |
| NYT Connections (extended) | 80% | 69.9% |
| ARC-AGI-1 | 87.5% | 65.3% |
| CritPt | 19.7% | 3.1% |
| Chess Puzzles | 40% | 12% |
| LMArena Hard Prompts | 1447 | 1453 |
| Epoch Capabilities Index | 156.44 | 148.03 |
| Kagi LLM Benchmark | — | 78.5% |
| EnigmaEval | — | 3.4% |
| Thematic Generalization | — | 69.4% |
| EBR-Bench | 30.5% | — |
| Mystery Game Puzzles | 34% | — |
| DTBench | 97.3% | — |
| LMCA | 48.5% | — |
Math Grok 4.6 leads
Grok 4.6: 67.0 (#24), Kimi K2.5: 51.8 (#53)
| Benchmark | Grok 4.6 | Kimi K2.5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 99.2% | 92.2% |
| LMArena Math | 1423 | 1470 |
| FrontierMath (Tiers 1-3) | 66% | — |
| FrontierMath Tier 4 | 31.7% | — |
| MathArena Final-Answer Competitions | — | 62.3% |
| ProofBench | 51% | — |
| FrontierMath (Feb 2025 set) | — | 27.9% |
| FrontierMath Tier 4 (v1) | — | 4.2% |
Knowledge Grok 4.6 leads
Grok 4.6: 63.3 (#20), Kimi K2.5: 53.6 (#56)
| Benchmark | Grok 4.6 | Kimi K2.5 |
|---|---|---|
| GPQA Diamond | 94% | 87.6% |
| SimpleQA Verified | 49.3% | 34.3% |
| LMArena Expert | 1467 | 1466 |
| Humanity's Last Exam | — | 24.4% |
| Vectara Hallucination Rate | — | 14.2% |
Multimodal Grok 4.6 leads
Grok 4.6: 43.6 (#23), Kimi K2.5: 41.1 (#39)
| Benchmark | Grok 4.6 | Kimi K2.5 |
|---|---|---|
| LMArena Vision | 1263 | 1269 |
| LMArena Document | 1452 | 1430 |
| Blueprint-Bench 2 | 33.2% | — |
| Furniture Assembly | 40% | — |
Multilingual Too close to call
Grok 4.6: 53.0 (#74), Kimi K2.5: 53.9 (#53)
| Benchmark | Grok 4.6 | Kimi K2.5 |
|---|---|---|
| LMArena Non-English | 1420 | 1433 |
| LMArena Chinese | 1480 | 1495 |
| LMArena French | 1461 | 1454 |
| LMArena German | 1431 | 1441 |
| LMArena Japanese | 1376 | 1421 |
| LMArena Korean | 1397 | 1410 |
| LMArena Russian | 1422 | 1435 |
| LMArena Spanish | 1404 | 1450 |
Instruction Following Too close to call
Grok 4.6: 75.4 (#63), Kimi K2.5: 75.3 (#64)
| Benchmark | Grok 4.6 | Kimi K2.5 |
|---|---|---|
| LMArena Instruction Following | 1431 | 1431 |
Long Context Kimi K2.5 leads
Grok 4.6: 44.5 (#66), Kimi K2.5: 52.1 (#7)
| Benchmark | Grok 4.6 | Kimi K2.5 |
|---|---|---|
| LMArena Longer Query | 1454 | 1445 |
| Fiction.LiveBench | — | 86.1% |
| CL-bench | — | 19.3% |
| CL-bench Life | — | 13.2% |
Writing & Preference Kimi K2.5 leads
Grok 4.6: 62.3 (#80), Kimi K2.5: 65.1 (#53)
| Benchmark | Grok 4.6 | Kimi K2.5 |
|---|---|---|
| LMArena Text | 1428 | 1445 |
| LMArena Creative Writing | 1428 | 1423 |
| LMArena Multi-Turn | 1425 | 1444 |
| EQ-Bench Creative Writing | — | 1579 |
Frequently asked questions
Is Grok 4.6 better than Kimi K2.5?
Grok 4.6 is the stronger model overall, scoring 56.9 to 48.1 on the Noometry Index. Kimi K2.5 costs 3.3× less per token, which makes it the better buy when Grok 4.6's lead doesn't matter for your workload.
Which is cheaper, Grok 4.6 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; Grok 4.6 lists at $2 and $6.
Is Grok 4.6 or Kimi K2.5 better for coding?
Grok 4.6 scores higher on coding benchmarks: 58.5 versus 48.8 in the Noometry coding category.
Which has the bigger context window?
Grok 4.6 does, with 500K tokens against 262K.
How many benchmarks do Grok 4.6 and Kimi K2.5 share?
34 benchmarks have published results for both models. Grok 4.6 has 49 scored results on Noometry and Kimi K2.5 has 51.