# Gemini 3 Flash Preview vs GPT-5.5

> GPT-5.5 is the stronger model overall, scoring 63.4 to 52.3 on the Noometry Index. Gemini 3 Flash Preview costs 10× less per token, which makes it the better buy when GPT-5.5's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/gemini-3-flash-preview-vs-gpt-5-5
- Last updated: 2026-10-10
- Shared benchmarks: 51

## Summary

- They share 51 benchmarks with published results for both. Gemini 3 Flash Preview scores higher in 0 categories and GPT-5.5 in 10 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.5 leads 81.5 to 51.4.
- The biggest single-benchmark swing is FrontierMath Tier 4: 17.1% for Gemini 3 Flash Preview and 72.5% for GPT-5.5.
- Gemini 3 Flash Preview is cheaper at $0.50 / $3 per million input/output tokens, against $5 / $30 for GPT-5.5.
- GPT-5.5 accepts more context: 1.05M tokens versus 1.05M.

## Snapshot

| | Gemini 3 Flash Preview | GPT-5.5 |
|---|---|---|
| Provider | Google | OpenAI |
| Noometry Index | 52.3 | 63.4 |
| Rank | 33 | 8 |
| Context | 1.05M | 1.05M |
| Input $/M | $0.50 | $5 |
| Output $/M | $3 | $30 |
| Weights | Proprietary | Proprietary |

## Coding

- Gemini 3 Flash Preview: 51.5 (#35)
- GPT-5.5: 58.3 (#13)

| Benchmark | Gemini 3 Flash Preview | GPT-5.5 |
|---|---|---|
| SWE-bench Verified | 75.4% | 80.6% |
| LMArena WebDev | 1439 | 1512 |
| GSO | 9.8% | 40.2% |
| WeirdML | 61.6% | 84.9% |
| LMArena Coding | 1460 | 1494 |
| ALE-Bench | 1,367 | 1,943 |
| DeepSWE | — | 67% |
| FrontierCode | — | 43% |
| SWE-bench Verified (bash only) | 75.8% | — |
| SWE-bench Multilingual | 72.7% | — |
| SciCode | — | 56.1% |
| MirrorCode | — | 10% |

## Agentic & Tool Use

- Gemini 3 Flash Preview: 38.6 (#27)
- GPT-5.5: 50.7 (#6)

| Benchmark | Gemini 3 Flash Preview | GPT-5.5 |
|---|---|---|
| Terminal-Bench | 64.3% | 84.7% |
| τ²-bench Banking | 27.3% | 44.6% |
| DeepResearch Bench | 49.8% | 54% |
| GDP.pdf | 10% | 26% |
| LMArena Search | 1198 | 1242 |
| Vending-Bench 2 | 3,635 | 7,524 |
| APEX-Agents | — | 55.1% |
| OSWorld 2.0 | — | 13% |
| Remote Labor Index | — | 6.3% |
| τ²-bench Airline | 82.5% | — |
| τ²-bench Retail | 76.8% | — |
| τ²-bench Telecom | 91.2% | — |
| PostTrainBench | — | 27.2% |
| BALROG | 48.1% | — |
| ExploitBench | — | 47.4% |
| GBAEval | — | 53.2% |

## Reasoning

- Gemini 3 Flash Preview: 49.2 (#35)
- GPT-5.5: 72.7 (#8)

| Benchmark | Gemini 3 Flash Preview | GPT-5.5 |
|---|---|---|
| ARC-AGI-2 | 33.6% | 85% |
| SimpleBench | 61.1% | 69% |
| NYT Connections (extended) | 83.1% | 96.2% |
| ARC-AGI-1 | 84.7% | 95% |
| Chess Puzzles | 40% | 54% |
| LMArena Hard Prompts | 1465 | 1489 |
| Mystery Game Puzzles | 26% | 56% |
| DTBench | 89.1% | 96% |
| LMCA | 43.1% | 54.3% |
| Epoch Capabilities Index | 151.83 | 159.25 |
| ForecastBench | 58.5 | 60.6 |
| Kagi LLM Benchmark | — | 88.8% |
| CritPt | — | 27.1% |
| EBR-Bench | — | 34.3% |
| Surface Evolver Bench | — | 88.1% |
| Bench to the Future 3 | — | 0.14 |

## Math

- Gemini 3 Flash Preview: 51.4 (#42)
- GPT-5.5: 81.5 (#9)

| Benchmark | Gemini 3 Flash Preview | GPT-5.5 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 51.2% | 85.3% |
| FrontierMath Tier 4 | 17.1% | 72.5% |
| MathArena Final-Answer Competitions | 67.6% | 94.3% |
| OTIS Mock AIME 2024-2025 | 95.6% | 100% |
| ProofBench | 15% | 50% |
| LMArena Math | 1472 | 1486 |
| FrontierMath (Feb 2025 set) | 35.6% | 51.7% |
| FrontierMath Tier 4 (v1) | 4.2% | 35.4% |
| FrontierMath Erdős | — | 0% |

## Knowledge

- Gemini 3 Flash Preview: 58.8 (#30)
- GPT-5.5: 64.4 (#14)

| Benchmark | Gemini 3 Flash Preview | GPT-5.5 |
|---|---|---|
| GPQA Diamond | 89.4% | 94% |
| SimpleQA Verified | 66.8% | 63% |
| Vectara Hallucination Rate | 13.5% | 9.3% |
| LMArena Expert | 1461 | 1508 |

## Multimodal

- Gemini 3 Flash Preview: 45.3 (#15)
- GPT-5.5: 47.1 (#11)

| Benchmark | Gemini 3 Flash Preview | GPT-5.5 |
|---|---|---|
| LMArena Vision | 1285 | 1297 |
| Blueprint-Bench 2 | 0% | 36.2% |
| LMArena Document | 1413 | 1486 |
| GeoBench | 88% | — |
| VPCT | 72.6% | — |
| Furniture Assembly | — | 44.2% |

## Multilingual

- Gemini 3 Flash Preview: 55.9 (#26)
- GPT-5.5: 56.5 (#19)

| Benchmark | Gemini 3 Flash Preview | GPT-5.5 |
|---|---|---|
| LMArena Non-English | 1458 | 1467 |
| LMArena Chinese | 1510 | 1535 |
| LMArena French | 1478 | 1486 |
| LMArena German | 1497 | 1478 |
| LMArena Japanese | 1488 | 1498 |
| LMArena Korean | 1443 | 1461 |
| LMArena Russian | 1480 | 1473 |
| LMArena Spanish | 1470 | 1469 |

## Instruction Following

- Gemini 3 Flash Preview: 75.8 (#52)
- GPT-5.5: 77.7 (#18)

| Benchmark | Gemini 3 Flash Preview | GPT-5.5 |
|---|---|---|
| LMArena Instruction Following | 1437 | 1479 |

## Long Context

- Gemini 3 Flash Preview: 44.4 (#63)
- GPT-5.5: 48.3 (#12)

| Benchmark | Gemini 3 Flash Preview | GPT-5.5 |
|---|---|---|
| LMArena Longer Query | 1452 | 1484 |
| CL-bench Life | — | 22.2% |

## Writing & Preference

- Gemini 3 Flash Preview: 65.6 (#44)
- GPT-5.5: 72.8 (#13)

| Benchmark | Gemini 3 Flash Preview | GPT-5.5 |
|---|---|---|
| LMArena Text | 1466 | 1471 |
| LMArena Creative Writing | 1457 | 1455 |
| LMArena Multi-Turn | 1471 | 1477 |
| EQ-Bench Creative Writing | — | 1844 |
| EQ-Bench 4 | — | 1315 |

## FAQ

### Is Gemini 3 Flash Preview better than GPT-5.5?

GPT-5.5 is the stronger model overall, scoring 63.4 to 52.3 on the Noometry Index. Gemini 3 Flash Preview costs 10× less per token, which makes it the better buy when GPT-5.5's lead doesn't matter for your workload.

### Which is cheaper, Gemini 3 Flash Preview or GPT-5.5?

Gemini 3 Flash Preview is cheaper. It lists at $0.50 per million input tokens and $3 per million output tokens; GPT-5.5 lists at $5 and $30.

### Is Gemini 3 Flash Preview or GPT-5.5 better for coding?

GPT-5.5 scores higher on coding benchmarks: 58.3 versus 51.5 in the Noometry coding category.

### Which has the bigger context window?

GPT-5.5 does, with 1.05M tokens against 1.05M.

### How many benchmarks do Gemini 3 Flash Preview and GPT-5.5 share?

51 benchmarks have published results for both models. Gemini 3 Flash Preview has 59 scored results on Noometry and GPT-5.5 has 71.