# Gemini 3 Flash Preview vs GPT-5

> Gemini 3 Flash Preview is the stronger model overall, scoring 52.3 to 50.9 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gemini-3-flash-preview-vs-gpt-5
- Last updated: 2026-10-11
- Shared benchmarks: 48

## Summary

- They share 48 benchmarks with published results for both. Gemini 3 Flash Preview scores higher in 7 categories and GPT-5 in 3 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in long context, where GPT-5 leads 69.5 to 44.4.
- The biggest single-benchmark swing is ARC-AGI-2: 33.6% for Gemini 3 Flash Preview and 9.9% for GPT-5.
- Gemini 3 Flash Preview is cheaper at $0.50 / $3 per million input/output tokens, against $1.25 / $10 for GPT-5.
- Gemini 3 Flash Preview accepts more context: 1.05M tokens versus 400K.

## Snapshot

| | Gemini 3 Flash Preview | GPT-5 |
|---|---|---|
| Provider | Google | OpenAI |
| Noometry Index | 52.3 | 50.9 |
| Rank | 40 | 45 |
| Context | 1.05M | 400K |
| Input $/M | $0.50 | $1.25 |
| Output $/M | $3 | $10 |
| Weights | Proprietary | Proprietary |

## Coding

- Gemini 3 Flash Preview: 50.9 (#42)
- GPT-5: 50.3 (#47)

| Benchmark | Gemini 3 Flash Preview | GPT-5 |
|---|---|---|
| SWE-bench Verified | 75.4% | 73.6% |
| SWE-bench Verified (bash only) | 75.8% | 65% |
| LMArena WebDev | 1439 | 1418 |
| GSO | 9.8% | 6.9% |
| WeirdML | 61.6% | 60.7% |
| LMArena Coding | 1460 | 1436 |
| ALE-Bench | 1,367 | 1,162 |
| Aider Polyglot | — | 88% |
| SWE-bench Multilingual | 72.7% | — |
| SciCode | — | 42.9% |
| AlgoTune | — | 1.67 |

## Agentic & Tool Use

- Gemini 3 Flash Preview: 38.7 (#29)
- GPT-5: 33.1 (#56)

| Benchmark | Gemini 3 Flash Preview | GPT-5 |
|---|---|---|
| Terminal-Bench | 64.3% | 49.6% |
| DeepResearch Bench | 49.8% | 49.6% |
| BALROG | 48.1% | 32.8% |
| LMArena Search | 1198 | 1133 |
| GDPval | — | 34.8% |
| Remote Labor Index | — | 1.7% |
| τ²-bench Airline | 82.5% | — |
| τ²-bench Banking | 27.3% | — |
| τ²-bench Retail | 76.8% | — |
| τ²-bench Telecom | 91.2% | — |
| GDP.pdf | 10% | — |
| METR Time Horizons | — | 69.6% |
| Vending-Bench 2 | 3,635 | — |

## Reasoning

- Gemini 3 Flash Preview: 49.2 (#37)
- GPT-5: 38.3 (#64)

| Benchmark | Gemini 3 Flash Preview | GPT-5 |
|---|---|---|
| ARC-AGI-2 | 33.6% | 9.9% |
| SimpleBench | 61.1% | 56.7% |
| ARC-AGI-1 | 84.7% | 65.7% |
| Chess Puzzles | 40% | 37% |
| LMArena Hard Prompts | 1465 | 1416 |
| Mystery Game Puzzles | 26% | 23% |
| DTBench | 89.1% | 90.7% |
| LMCA | 43.1% | 40% |
| Epoch Capabilities Index | 151.8 | 150 |
| ForecastBench | 58.5 | 61.4 |
| Kagi LLM Benchmark | — | 72.7% |
| NYT Connections (extended) | 83.1% | — |
| CritPt | — | 12.6% |
| EnigmaEval | — | 10.5% |
| EBR-Bench | — | 12.7% |

## Math

- Gemini 3 Flash Preview: 51.7 (#55)
- GPT-5: 55.0 (#44)

| Benchmark | Gemini 3 Flash Preview | GPT-5 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 51.2% | 55.4% |
| FrontierMath Tier 4 | 17.1% | 22% |
| OTIS Mock AIME 2024-2025 | 95.6% | 91.4% |
| ProofBench | 15% | 18% |
| LMArena Math | 1473 | 1407 |
| FrontierMath (Feb 2025 set) | 35.6% | 32.4% |
| FrontierMath Tier 4 (v1) | 4.2% | 12.5% |
| MathArena Final-Answer Competitions | 67.6% | — |
| Omni-MATH | — | 64.7% |
| MATH Level 5 | — | 98.1% |

## Knowledge

- Gemini 3 Flash Preview: 58.8 (#33)
- GPT-5: 56.6 (#43)

| Benchmark | Gemini 3 Flash Preview | GPT-5 |
|---|---|---|
| GPQA Diamond | 89.4% | 86.2% |
| SimpleQA Verified | 66.8% | 50.1% |
| Vectara Hallucination Rate | 13.5% | 14.7% |
| LMArena Expert | 1462 | 1419 |
| Humanity's Last Exam | — | 25.3% |
| MMLU-Pro | — | 86.3% |
| Confabulations | — | 10.3% |
| GPQA (HELM) | — | 79.2% |

## Multimodal

- Gemini 3 Flash Preview: 45.5 (#16)
- GPT-5: 46.8 (#13)

| Benchmark | Gemini 3 Flash Preview | GPT-5 |
|---|---|---|
| LMArena Vision | 1285 | 1232 |
| GeoBench | 88% | 81% |
| VPCT | 72.6% | 66% |
| Blueprint-Bench 2 | 0% | — |
| LMArena Document | 1413 | — |

## Multilingual

- Gemini 3 Flash Preview: 55.7 (#27)
- GPT-5: 51.4 (#110)

| Benchmark | Gemini 3 Flash Preview | GPT-5 |
|---|---|---|
| LMArena Non-English | 1458 | 1397 |
| LMArena Chinese | 1511 | 1422 |
| LMArena French | 1477 | 1410 |
| LMArena German | 1497 | 1416 |
| LMArena Japanese | 1489 | 1409 |
| LMArena Korean | 1443 | 1360 |
| LMArena Russian | 1480 | 1406 |
| LMArena Spanish | 1469 | 1399 |

## Instruction Following

- Gemini 3 Flash Preview: 75.7 (#56)
- GPT-5: 73.8 (#113)

| Benchmark | Gemini 3 Flash Preview | GPT-5 |
|---|---|---|
| LMArena Instruction Following | 1437 | 1388 |
| IFEval | — | 87.5% |

## Long Context

- Gemini 3 Flash Preview: 44.4 (#67)
- GPT-5: 69.5 (#2)

| Benchmark | Gemini 3 Flash Preview | GPT-5 |
|---|---|---|
| LMArena Longer Query | 1452 | 1399 |
| Fiction.LiveBench | — | 97.2% |

## Writing & Preference

- Gemini 3 Flash Preview: 65.5 (#45)
- GPT-5: 63.4 (#65)

| Benchmark | Gemini 3 Flash Preview | GPT-5 |
|---|---|---|
| LMArena Text | 1466 | 1406 |
| LMArena Creative Writing | 1457 | 1365 |
| LMArena Multi-Turn | 1471 | 1426 |
| Short-Story Creative Writing | — | 86% |
| EQ-Bench Creative Writing | — | 1627 |
| WildBench | — | 85.7% |

## FAQ

### Is Gemini 3 Flash Preview better than GPT-5?

Gemini 3 Flash Preview is the stronger model overall, scoring 52.3 to 50.9 on the Noometry Index.

### Which is cheaper, Gemini 3 Flash Preview or GPT-5?

Gemini 3 Flash Preview is cheaper. It lists at $0.50 per million input tokens and $3 per million output tokens; GPT-5 lists at $1.25 and $10.

### Is Gemini 3 Flash Preview or GPT-5 better for coding?

They score almost the same on coding (50.9 vs 50.3); test both on your own repository before choosing.

### Which has the bigger context window?

Gemini 3 Flash Preview does, with 1.05M tokens against 400K.

### How many benchmarks do Gemini 3 Flash Preview and GPT-5 share?

48 benchmarks have published results for both models. Gemini 3 Flash Preview has 59 scored results on Noometry and GPT-5 has 69.