Model comparison
Gemini 3 Flash Preview vs Kimi K3
Kimi K3 is the stronger model overall, scoring 59.5 to 52.3 on the Noometry Index. Gemini 3 Flash Preview costs 5.3× less per token, which makes it the better buy when Kimi K3's lead doesn't matter for your workload.
Last verified . 41 shared benchmarks.
GoogleSummary
- They share 41 benchmarks with published results for both. Gemini 3 Flash Preview scores higher in 1 category and Kimi K3 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K3 leads 74.2 to 51.7.
- The biggest single-benchmark swing is ProofBench: 15% for Gemini 3 Flash Preview and 87% for Kimi K3.
- Gemini 3 Flash Preview is cheaper at $0.50 / $3 per million input/output tokens, against $3 / $15 for Kimi K3.
- Kimi K3 has downloadable open weights; the other is API-only.
Side by side
| Gemini 3 Flash Preview | Kimi K3 | |
|---|---|---|
| Provider | Moonshot AI | |
| Noometry Index | 52.3 | 59.5 |
| Released | 2025-12-17 | 2026-07-16 |
| Weights | Proprietary | Open |
| Context window | 1.05M | 1.05M |
| Max output | 66K | 1.05M |
| Input $ / M tokens | $0.50 | $3 |
| Output $ / M tokens | $3 | $15 |
| Results tracked | 59 | 53 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K3 leads
Gemini 3 Flash Preview: 50.9 (#42), Kimi K3: 61.0 (#10)
| Benchmark | Gemini 3 Flash Preview | Kimi K3 |
|---|---|---|
| LMArena WebDev | 1439 | 1654 |
| WeirdML | 61.6% | 82.6% |
| LMArena Coding | 1460 | 1508 |
| ALE-Bench | 1,367 | 1,524 |
| SWE-bench Verified | 75.4% | — |
| DeepSWE | — | 68.5% |
| FrontierCode | — | 44.2% |
| SWE-bench Verified (bash only) | 75.8% | — |
| SWE-bench Multilingual | 72.7% | — |
| FrontierSWE | — | 25.9% |
| SciCode | — | 59.5% |
| GSO | 9.8% | — |
Agentic & Tool Use Kimi K3 leads
Gemini 3 Flash Preview: 38.7 (#29), Kimi K3: 41.8 (#20)
| Benchmark | Gemini 3 Flash Preview | Kimi K3 |
|---|---|---|
| τ²-bench Banking | 27.3% | 37.1% |
| GDP.pdf | 10% | 19% |
| Vending-Bench 2 | 3,635 | 5,165 |
| Terminal-Bench | 64.3% | — |
| APEX-Agents | — | 50.6% |
| τ²-bench Airline | 82.5% | — |
| τ²-bench Retail | 76.8% | — |
| τ²-bench Telecom | 91.2% | — |
| DeepResearch Bench | 49.8% | — |
| PostTrainBench | — | 32% |
| BALROG | 48.1% | — |
| GBAEval | — | 48.3% |
| LMArena Search | 1198 | — |
Reasoning Kimi K3 leads
Gemini 3 Flash Preview: 49.2 (#37), Kimi K3: 63.0 (#17)
| Benchmark | Gemini 3 Flash Preview | Kimi K3 |
|---|---|---|
| ARC-AGI-2 | 33.6% | 60.4% |
| SimpleBench | 61.1% | 60.7% |
| NYT Connections (extended) | 83.1% | 93.6% |
| ARC-AGI-1 | 84.7% | 94.5% |
| Chess Puzzles | 40% | 39% |
| LMArena Hard Prompts | 1465 | 1496 |
| Mystery Game Puzzles | 26% | 26% |
| DTBench | 89.1% | 91.2% |
| LMCA | 43.1% | 52.7% |
| Epoch Capabilities Index | 151.8 | 157.45 |
| ForecastBench | 58.5 | 61.1 |
| CritPt | — | 23.4% |
| Surface Evolver Bench | — | 95% |
Math Kimi K3 leads
Gemini 3 Flash Preview: 51.7 (#55), Kimi K3: 74.2 (#16)
| Benchmark | Gemini 3 Flash Preview | Kimi K3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 51.2% | 72.2% |
| FrontierMath Tier 4 | 17.1% | 39% |
| MathArena Final-Answer Competitions | 67.6% | 87.8% |
| OTIS Mock AIME 2024-2025 | 95.6% | 97.2% |
| ProofBench | 15% | 87% |
| LMArena Math | 1473 | 1491 |
| FrontierMath (Feb 2025 set) | 35.6% | — |
| FrontierMath Tier 4 (v1) | 4.2% | — |
Knowledge Kimi K3 leads
Gemini 3 Flash Preview: 58.8 (#33), Kimi K3: 63.2 (#21)
| Benchmark | Gemini 3 Flash Preview | Kimi K3 |
|---|---|---|
| GPQA Diamond | 89.4% | 93.1% |
| SimpleQA Verified | 66.8% | 50.6% |
| LMArena Expert | 1462 | 1521 |
| Vectara Hallucination Rate | 13.5% | — |
Multimodal Gemini 3 Flash Preview leads
Gemini 3 Flash Preview: 45.5 (#16), Kimi K3: 37.8 (#70)
| Benchmark | Gemini 3 Flash Preview | Kimi K3 |
|---|---|---|
| Blueprint-Bench 2 | 0% | 29.5% |
| LMArena Vision | 1285 | — |
| GeoBench | 88% | — |
| VPCT | 72.6% | — |
| Furniture Assembly | — | 34.2% |
| LMArena Document | 1413 | — |
Multilingual Too close to call
Gemini 3 Flash Preview: 55.7 (#27), Kimi K3: 56.3 (#21)
| Benchmark | Gemini 3 Flash Preview | Kimi K3 |
|---|---|---|
| LMArena Non-English | 1458 | 1466 |
| LMArena Chinese | 1511 | 1529 |
| LMArena French | 1477 | 1491 |
| LMArena German | 1497 | 1488 |
| LMArena Japanese | 1489 | 1487 |
| LMArena Korean | 1443 | 1458 |
| LMArena Russian | 1480 | 1482 |
| LMArena Spanish | 1469 | 1472 |
Instruction Following Kimi K3 leads
Gemini 3 Flash Preview: 75.7 (#56), Kimi K3: 77.7 (#14)
| Benchmark | Gemini 3 Flash Preview | Kimi K3 |
|---|---|---|
| LMArena Instruction Following | 1437 | 1483 |
Long Context Kimi K3 leads
Gemini 3 Flash Preview: 44.4 (#67), Kimi K3: 45.8 (#29)
| Benchmark | Gemini 3 Flash Preview | Kimi K3 |
|---|---|---|
| LMArena Longer Query | 1452 | 1494 |
Writing & Preference Kimi K3 leads
Gemini 3 Flash Preview: 65.5 (#45), Kimi K3: 76.6 (#4)
| Benchmark | Gemini 3 Flash Preview | Kimi K3 |
|---|---|---|
| LMArena Text | 1466 | 1476 |
| LMArena Creative Writing | 1457 | 1454 |
| LMArena Multi-Turn | 1471 | 1488 |
| EQ-Bench Creative Writing | — | 2082 |
| EQ-Bench 4 | — | 1339 |
Frequently asked questions
Is Gemini 3 Flash Preview better than Kimi K3?
Kimi K3 is the stronger model overall, scoring 59.5 to 52.3 on the Noometry Index. Gemini 3 Flash Preview costs 5.3× less per token, which makes it the better buy when Kimi K3's lead doesn't matter for your workload.
Which is cheaper, Gemini 3 Flash Preview or Kimi K3?
Gemini 3 Flash Preview is cheaper. It lists at $0.50 per million input tokens and $3 per million output tokens; Kimi K3 lists at $3 and $15.
Is Gemini 3 Flash Preview or Kimi K3 better for coding?
Kimi K3 scores higher on coding benchmarks: 61.0 versus 50.9 in the Noometry coding category.
Which has the bigger context window?
Both accept 1.05M tokens.
How many benchmarks do Gemini 3 Flash Preview and Kimi K3 share?
41 benchmarks have published results for both models. Gemini 3 Flash Preview has 59 scored results on Noometry and Kimi K3 has 53.