Model comparison
Gemini 3 Flash Preview vs Gemini 3 Pro
Gemini 3 Pro is the stronger model overall, scoring 54.8 to 52.3 on the Noometry Index.
Last verified . 51 shared benchmarks.
GoogleSummary
- They share 51 benchmarks with published results for both. Gemini 3 Flash Preview scores higher in 2 categories and Gemini 3 Pro in 8 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in multimodal, where Gemini 3 Pro leads 57.6 to 45.5.
- The biggest single-benchmark swing is VPCT: 72.6% for Gemini 3 Flash Preview and 91% for Gemini 3 Pro.
Side by side
| Gemini 3 Flash Preview | Gemini 3 Pro | |
|---|---|---|
| Provider | ||
| Noometry Index | 52.3 | 54.8 |
| Released | 2025-12-17 | 2025-11-18 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | — |
| Max output | 66K | — |
| Input $ / M tokens | $0.50 | — |
| Output $ / M tokens | $3 | — |
| Results tracked | 59 | 67 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Gemini 3 Flash Preview: 50.9 (#42), Gemini 3 Pro: 51.6 (#39)
| Benchmark | Gemini 3 Flash Preview | Gemini 3 Pro |
|---|---|---|
| SWE-bench Verified | 75.4% | 72.9% |
| SWE-bench Verified (bash only) | 75.8% | 74.2% |
| LMArena WebDev | 1439 | 1440 |
| SWE-bench Multilingual | 72.7% | 68.7% |
| GSO | 9.8% | 18.6% |
| WeirdML | 61.6% | 69.9% |
| LMArena Coding | 1460 | 1481 |
| ALE-Bench | 1,367 | 1,177 |
| AlgoTune | — | 1.83 |
Agentic & Tool Use Gemini 3 Pro leads
Gemini 3 Flash Preview: 38.7 (#29), Gemini 3 Pro: 40.6 (#23)
| Benchmark | Gemini 3 Flash Preview | Gemini 3 Pro |
|---|---|---|
| Terminal-Bench | 64.3% | 69.4% |
| τ²-bench Airline | 82.5% | 80.5% |
| τ²-bench Banking | 27.3% | 18% |
| τ²-bench Retail | 76.8% | 75.9% |
| τ²-bench Telecom | 91.2% | 91% |
| DeepResearch Bench | 49.8% | 46.3% |
| BALROG | 48.1% | 58.1% |
| LMArena Search | 1198 | 1207 |
| Vending-Bench 2 | 3,635 | 5,478 |
| Berkeley Function Calling Leaderboard | — | 72.5% |
| GDPval | — | 40.3% |
| Remote Labor Index | — | 1.3% |
| GDP.pdf | 10% | — |
| METR Time Horizons | — | 71% |
Reasoning Gemini 3 Pro leads
Gemini 3 Flash Preview: 49.2 (#37), Gemini 3 Pro: 52.5 (#31)
| Benchmark | Gemini 3 Flash Preview | Gemini 3 Pro |
|---|---|---|
| ARC-AGI-2 | 33.6% | 31.1% |
| SimpleBench | 61.1% | 76.4% |
| NYT Connections (extended) | 83.1% | 94.4% |
| ARC-AGI-1 | 84.7% | 75% |
| Chess Puzzles | 40% | 31% |
| LMArena Hard Prompts | 1465 | 1480 |
| Epoch Capabilities Index | 151.8 | 152.92 |
| ForecastBench | 58.5 | 61.2 |
| Kagi LLM Benchmark | — | 80.1% |
| CritPt | — | 6.9% |
| EnigmaEval | — | 18.2% |
| Mystery Game Puzzles | 26% | — |
| DTBench | 89.1% | — |
| LMCA | 43.1% | — |
Math Gemini 3 Flash Preview leads
Gemini 3 Flash Preview: 51.7 (#55), Gemini 3 Pro: 49.9 (#59)
| Benchmark | Gemini 3 Flash Preview | Gemini 3 Pro |
|---|---|---|
| MathArena Final-Answer Competitions | 67.6% | 67% |
| OTIS Mock AIME 2024-2025 | 95.6% | 91.4% |
| ProofBench | 15% | 20% |
| LMArena Math | 1473 | 1476 |
| FrontierMath (Feb 2025 set) | 35.6% | 37.6% |
| FrontierMath Tier 4 (v1) | 4.2% | 18.8% |
| FrontierMath (Tiers 1-3) | 51.2% | — |
| FrontierMath Tier 4 | 17.1% | — |
| Omni-MATH | — | 55.5% |
Knowledge Gemini 3 Pro leads
Gemini 3 Flash Preview: 58.8 (#33), Gemini 3 Pro: 64.4 (#16)
| Benchmark | Gemini 3 Flash Preview | Gemini 3 Pro |
|---|---|---|
| GPQA Diamond | 89.4% | 92.6% |
| Vectara Hallucination Rate | 13.5% | 13.6% |
| LMArena Expert | 1462 | 1475 |
| Humanity's Last Exam | — | 37.5% |
| SimpleQA Verified | 66.8% | — |
| MMLU-Pro | — | 90.3% |
| GPQA (HELM) | — | 80.3% |
Multimodal Gemini 3 Pro leads
Gemini 3 Flash Preview: 45.5 (#16), Gemini 3 Pro: 57.6 (#2)
| Benchmark | Gemini 3 Flash Preview | Gemini 3 Pro |
|---|---|---|
| LMArena Vision | 1285 | 1305 |
| GeoBench | 88% | 84% |
| VPCT | 72.6% | 91% |
| LMArena Document | 1413 | 1434 |
| Blueprint-Bench 2 | 0% | — |
Multilingual Gemini 3 Pro leads
Gemini 3 Flash Preview: 55.7 (#27), Gemini 3 Pro: 56.9 (#16)
| Benchmark | Gemini 3 Flash Preview | Gemini 3 Pro |
|---|---|---|
| LMArena Non-English | 1458 | 1474 |
| LMArena Chinese | 1511 | 1523 |
| LMArena French | 1477 | 1492 |
| LMArena German | 1497 | 1515 |
| LMArena Japanese | 1489 | 1510 |
| LMArena Korean | 1443 | 1448 |
| LMArena Russian | 1480 | 1493 |
| LMArena Spanish | 1469 | 1470 |
Instruction Following Too close to call
Gemini 3 Flash Preview: 75.7 (#56), Gemini 3 Pro: 76.3 (#45)
| Benchmark | Gemini 3 Flash Preview | Gemini 3 Pro |
|---|---|---|
| LMArena Instruction Following | 1437 | 1458 |
| IFEval | — | 87.7% |
Long Context Too close to call
Gemini 3 Flash Preview: 44.4 (#67), Gemini 3 Pro: 44.0 (#79)
| Benchmark | Gemini 3 Flash Preview | Gemini 3 Pro |
|---|---|---|
| LMArena Longer Query | 1452 | 1471 |
| CL-bench | — | 15.8% |
Writing & Preference Too close to call
Gemini 3 Flash Preview: 65.5 (#45), Gemini 3 Pro: 66.4 (#35)
| Benchmark | Gemini 3 Flash Preview | Gemini 3 Pro |
|---|---|---|
| LMArena Text | 1466 | 1479 |
| LMArena Creative Writing | 1457 | 1482 |
| LMArena Multi-Turn | 1471 | 1484 |
| EQ-Bench Creative Writing | — | 1525 |
| WildBench | — | 85.9% |
Frequently asked questions
Is Gemini 3 Flash Preview better than Gemini 3 Pro?
Gemini 3 Pro is the stronger model overall, scoring 54.8 to 52.3 on the Noometry Index.
Is Gemini 3 Flash Preview or Gemini 3 Pro better for coding?
They score almost the same on coding (50.9 vs 51.6); test both on your own repository before choosing.
How many benchmarks do Gemini 3 Flash Preview and Gemini 3 Pro share?
51 benchmarks have published results for both models. Gemini 3 Flash Preview has 59 scored results on Noometry and Gemini 3 Pro has 67.