Model comparison
Gemma 3 12B vs Kimi K2 (Jul 2025)
Kimi K2 (Jul 2025) is the stronger model overall, scoring 41.2 to 32.1 on the Noometry Index. Gemma 3 12B costs 13× less per token, which makes it the better buy when Kimi K2 (Jul 2025)'s lead doesn't matter for your workload.
Last verified . 16 shared benchmarks.
GoogleSummary
- They share 16 benchmarks with published results for both. Gemma 3 12B scores higher in 0 categories and Kimi K2 (Jul 2025) in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K2 (Jul 2025) leads 42.7 to 22.3.
- The biggest single-benchmark swing is Berkeley Function Calling Leaderboard: 30.4% for Gemma 3 12B and 59.1% for Kimi K2 (Jul 2025).
- Gemma 3 12B is cheaper at $0.05 / $0.15 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 131K.
Side by side
| Gemma 3 12B | Kimi K2 (Jul 2025) | |
|---|---|---|
| Provider | Moonshot AI | |
| Noometry Index | 32.1 | 41.2 |
| Released | 2025-03-12 | 2025-07-12 |
| Weights | Open | Open |
| Context window | 131K | 262K |
| Max output | 8K | 262K |
| Input $ / M tokens | $0.05 | $0.57 |
| Output $ / M tokens | $0.15 | $2.30 |
| Results tracked | 24 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K2 (Jul 2025) leads
Gemma 3 12B: 31.7 (#280), Kimi K2 (Jul 2025): 42.4 (#102)
| Benchmark | Gemma 3 12B | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Coding | 1281 | 1399 |
| SWE-bench Verified (bash only) | — | 63.4% |
| Aider Polyglot | — | 59.1% |
| SciCode | 17.4% | — |
| GSO | — | 4.9% |
| WeirdML | — | 42.8% |
| ALE-Bench | — | 597.5 |
Agentic & Tool Use Kimi K2 (Jul 2025) leads
Gemma 3 12B: 25.5 (#108), Kimi K2 (Jul 2025): 32.4 (#64)
| Benchmark | Gemma 3 12B | Kimi K2 (Jul 2025) |
|---|---|---|
| Berkeley Function Calling Leaderboard | 30.4% | 59.1% |
| Terminal-Bench | — | 35.7% |
| METR Time Horizons | — | 59.2% |
Reasoning Kimi K2 (Jul 2025) leads
Gemma 3 12B: 15.7 (#313), Kimi K2 (Jul 2025): 23.3 (#179)
| Benchmark | Gemma 3 12B | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Hard Prompts | 1309 | 1384 |
| Epoch Capabilities Index | 123.5 | 146.01 |
| SimpleBench | — | 26.3% |
| Kagi LLM Benchmark | — | 64.4% |
| CritPt | 0% | — |
| Chess Puzzles | 0% | — |
| DTBench | 48.8% | — |
| LMCA | 4.5% | — |
| ForecastBench | — | 60.2 |
Math Kimi K2 (Jul 2025) leads
Gemma 3 12B: 22.3 (#279), Kimi K2 (Jul 2025): 42.7 (#83)
| Benchmark | Gemma 3 12B | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Math | 1307 | 1397 |
| OTIS Mock AIME 2024-2025 | 16.7% | — |
| Omni-MATH | — | 65.4% |
| FrontierMath (Feb 2025 set) | — | 21.4% |
| FrontierMath Tier 4 (v1) | — | 0% |
Knowledge Kimi K2 (Jul 2025) leads
Gemma 3 12B: 26.5 (#257), Kimi K2 (Jul 2025): 37.3 (#157)
| Benchmark | Gemma 3 12B | Kimi K2 (Jul 2025) |
|---|---|---|
| Vectara Hallucination Rate | 4.4% | 17.9% |
| LMArena Expert | 1248 | 1365 |
| GPQA Diamond | 39.5% | — |
| MMLU-Pro | — | 81.9% |
| Confabulations | — | 20.4% |
| GPQA (HELM) | — | 65.3% |
Multimodal Not comparable
Gemma 3 12B: —, Kimi K2 (Jul 2025): —
| Benchmark | Gemma 3 12B | Kimi K2 (Jul 2025) |
|---|---|---|
| MindCube | 46.7% | — |
Multilingual Kimi K2 (Jul 2025) leads
Gemma 3 12B: 45.7 (#165), Kimi K2 (Jul 2025): 49.6 (#130)
| Benchmark | Gemma 3 12B | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Non-English | 1318 | 1372 |
| LMArena German | 1370 | 1387 |
| LMArena Russian | 1335 | 1385 |
| LMArena Chinese | — | 1415 |
| LMArena French | — | 1379 |
| LMArena Japanese | — | 1349 |
| LMArena Korean | — | 1325 |
| LMArena Spanish | — | 1386 |
Instruction Following Kimi K2 (Jul 2025) leads
Gemma 3 12B: 68.6 (#186), Kimi K2 (Jul 2025): 71.1 (#156)
| Benchmark | Gemma 3 12B | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Instruction Following | 1299 | 1348 |
| IFEval | — | 85% |
Long Context Kimi K2 (Jul 2025) leads
Gemma 3 12B: 40.0 (#162), Kimi K2 (Jul 2025): 41.2 (#145)
| Benchmark | Gemma 3 12B | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Longer Query | 1317 | 1353 |
| Fiction.LiveBench | — | 66.7% |
| CL-bench | — | 17.6% |
Writing & Preference Kimi K2 (Jul 2025) leads
Gemma 3 12B: 47.5 (#209), Kimi K2 (Jul 2025): 62.3 (#78)
| Benchmark | Gemma 3 12B | Kimi K2 (Jul 2025) |
|---|---|---|
| LMArena Text | 1334 | 1380 |
| LMArena Creative Writing | 1331 | 1350 |
| EQ-Bench Creative Writing | 1126 | 1666 |
| LMArena Multi-Turn | 1334 | 1371 |
| Short-Story Creative Writing | — | 85.6% |
| WildBench | — | 86.2% |
Frequently asked questions
Is Gemma 3 12B better than Kimi K2 (Jul 2025)?
Kimi K2 (Jul 2025) is the stronger model overall, scoring 41.2 to 32.1 on the Noometry Index. Gemma 3 12B costs 13× less per token, which makes it the better buy when Kimi K2 (Jul 2025)'s lead doesn't matter for your workload.
Which is cheaper, Gemma 3 12B or Kimi K2 (Jul 2025)?
Gemma 3 12B is cheaper. It lists at $0.05 per million input tokens and $0.15 per million output tokens; Kimi K2 (Jul 2025) lists at $0.57 and $2.30.
Is Gemma 3 12B or Kimi K2 (Jul 2025) better for coding?
Kimi K2 (Jul 2025) scores higher on coding benchmarks: 42.4 versus 31.7 in the Noometry coding category.
Which has the bigger context window?
Kimi K2 (Jul 2025) does, with 262K tokens against 131K.
How many benchmarks do Gemma 3 12B and Kimi K2 (Jul 2025) share?
16 benchmarks have published results for both models. Gemma 3 12B has 24 scored results on Noometry and Kimi K2 (Jul 2025) has 42.