Model comparison
Gemma 4 31B IT vs Kimi K2 (Jul 2025)
Gemma 4 31B IT is the stronger model overall, scoring 43.5 to 41.2 on the Noometry Index.
Last verified . 20 shared benchmarks.
GoogleSummary
- They share 20 benchmarks with published results for both. Gemma 4 31B IT scores higher in 6 categories and Kimi K2 (Jul 2025) in 2 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in instruction following, where Gemma 4 31B IT leads 75.5 to 71.1.
- The biggest single-benchmark swing is Vectara Hallucination Rate: 7.4% for Gemma 4 31B IT and 17.9% for Kimi K2 (Jul 2025).
- Gemma 4 31B IT is cheaper at $0.09 / $0.34 per million input/output tokens, against $0.57 / $2.30 for Kimi K2 (Jul 2025).
Side by side
| Gemma 4 31B IT | Kimi K2 (Jul 2025) | |
|---|---|---|
| Provider | Moonshot AI | |
| Noometry Index | 43.5 | 41.2 |
| Released | 2026-04-02 | 2025-07-12 |
| Weights | Open | Open |
| Context window | 262K | 262K |
| Max output | 33K | 262K |
| Input $ / M tokens | $0.09 | $0.57 |
| Output $ / M tokens | $0.34 | $2.30 |
| Results tracked | 35 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Gemma 4 31B IT: 42.3 (#108), Kimi K2 (Jul 2025): 42.4 (#102)
| Benchmark | Gemma 4 31B IT | Kimi K2 (Jul 2025) |
|---|---|---|
| WeirdML | 52.3% | 42.8% |
| LMArena Coding | 1459 | 1399 |
| ALE-Bench | 925.5 | 597.5 |
| SWE-bench Verified (bash only) | — | 63.4% |
| Aider Polyglot | — | 59.1% |
| LMArena WebDev | 1366 | — |
| SciCode | 43.4% | — |
| GSO | — | 4.9% |
Agentic & Tool Use Not comparable
Gemma 4 31B IT: —, Kimi K2 (Jul 2025): 32.4 (#64)
| Benchmark | Gemma 4 31B IT | Kimi K2 (Jul 2025) |
|---|---|---|
| Terminal-Bench | — | 35.7% |
| Berkeley Function Calling Leaderboard | — | 59.1% |
| METR Time Horizons | — | 59.2% |
Reasoning Gemma 4 31B IT leads
Gemma 4 31B IT: 27.2 (#122), Kimi K2 (Jul 2025): 23.3 (#179)
| Benchmark | Gemma 4 31B IT | Kimi K2 (Jul 2025) |
|---|---|---|
| Kagi LLM Benchmark | 63.5% | 64.4% |
| LMArena Hard Prompts | 1448 | 1384 |
| Epoch Capabilities Index | 142.74 | 146.01 |
| SimpleBench | — | 26.3% |
| NYT Connections (extended) | 70.6% | — |
| CritPt | 1.4% | — |
| Chess Puzzles | 5% | — |
| Thematic Generalization | 53% | — |
| DTBench | 82.7% | — |
| LMCA | 39.3% | — |
| Surface Evolver Bench | 30.6% | — |
| ForecastBench | — | 60.2 |
Math Too close to call
Gemma 4 31B IT: 43.2 (#81), Kimi K2 (Jul 2025): 42.7 (#83)
| Benchmark | Gemma 4 31B IT | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Math | 1465 | 1397 |
| OTIS Mock AIME 2024-2025 | 73.3% | — |
| Omni-MATH | — | 65.4% |
| FrontierMath (Feb 2025 set) | — | 21.4% |
| FrontierMath Tier 4 (v1) | — | 0% |
Knowledge Too close to call
Gemma 4 31B IT: 37.9 (#151), Kimi K2 (Jul 2025): 37.3 (#157)
| Benchmark | Gemma 4 31B IT | Kimi K2 (Jul 2025) |
|---|---|---|
| Vectara Hallucination Rate | 7.4% | 17.9% |
| LMArena Expert | 1465 | 1365 |
| GPQA Diamond | 75.8% | — |
| SimpleQA Verified | 10.4% | — |
| MMLU-Pro | — | 81.9% |
| Confabulations | — | 20.4% |
| GPQA (HELM) | — | 65.3% |
Multimodal Not comparable
Gemma 4 31B IT: 41.6 (#34), Kimi K2 (Jul 2025): —
| Benchmark | Gemma 4 31B IT | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Vision | 1277 | — |
| LMArena Document | 1425 | — |
Multilingual Gemma 4 31B IT leads
Gemma 4 31B IT: 53.8 (#57), Kimi K2 (Jul 2025): 49.6 (#130)
| Benchmark | Gemma 4 31B IT | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Non-English | 1431 | 1372 |
| LMArena Chinese | 1476 | 1415 |
| LMArena French | 1435 | 1379 |
| LMArena Russian | 1460 | 1385 |
| LMArena Spanish | 1444 | 1386 |
| LMArena German | — | 1387 |
| LMArena Japanese | — | 1349 |
| LMArena Korean | — | 1325 |
Instruction Following Gemma 4 31B IT leads
Gemma 4 31B IT: 75.5 (#61), Kimi K2 (Jul 2025): 71.1 (#156)
| Benchmark | Gemma 4 31B IT | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Instruction Following | 1433 | 1348 |
| IFEval | — | 85% |
Long Context Gemma 4 31B IT leads
Gemma 4 31B IT: 44.2 (#71), Kimi K2 (Jul 2025): 41.2 (#145)
| Benchmark | Gemma 4 31B IT | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Longer Query | 1446 | 1353 |
| Fiction.LiveBench | — | 66.7% |
| CL-bench | — | 17.6% |
Writing & Preference Kimi K2 (Jul 2025) leads
Gemma 4 31B IT: 60.5 (#96), Kimi K2 (Jul 2025): 62.3 (#78)
| Benchmark | Gemma 4 31B IT | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Text | 1443 | 1380 |
| LMArena Creative Writing | 1415 | 1350 |
| EQ-Bench Creative Writing | 1368 | 1666 |
| LMArena Multi-Turn | 1452 | 1371 |
| Short-Story Creative Writing | — | 85.6% |
| WildBench | — | 86.2% |
| EQ-Bench 4 | 1120 | — |
Frequently asked questions
Is Gemma 4 31B IT better than Kimi K2 (Jul 2025)?
Gemma 4 31B IT is the stronger model overall, scoring 43.5 to 41.2 on the Noometry Index.
Which is cheaper, Gemma 4 31B IT or Kimi K2 (Jul 2025)?
Gemma 4 31B IT is cheaper. It lists at $0.09 per million input tokens and $0.34 per million output tokens; Kimi K2 (Jul 2025) lists at $0.57 and $2.30.
Is Gemma 4 31B IT or Kimi K2 (Jul 2025) better for coding?
They score almost the same on coding (42.3 vs 42.4); test both on your own repository before choosing.
Which has the bigger context window?
Both accept 262K tokens.
How many benchmarks do Gemma 4 31B IT and Kimi K2 (Jul 2025) share?
20 benchmarks have published results for both models. Gemma 4 31B IT has 35 scored results on Noometry and Kimi K2 (Jul 2025) has 42.