# Gemini 3.5 Flash vs GPT-5.4

> GPT-5.4 is the stronger model overall, scoring 59.5 to 54.1 on the Noometry Index. Gemini 3.5 Flash costs 1.7× less per token, which makes it the better buy when GPT-5.4's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/gemini-3-5-flash-vs-gpt-5-4
- Last updated: 2026-10-10
- Shared benchmarks: 51

## Summary

- They share 51 benchmarks with published results for both. Gemini 3.5 Flash scores higher in 4 categories and GPT-5.4 in 6 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in agentic & tool use, where GPT-5.4 leads 46.5 to 24.8.
- The biggest single-benchmark swing is GBAEval: 6.7% for Gemini 3.5 Flash and 45.1% for GPT-5.4.
- Gemini 3.5 Flash is cheaper at $1.50 / $9 per million input/output tokens, against $2.50 / $15 for GPT-5.4.
- GPT-5.4 accepts more context: 1.05M tokens versus 1.05M.

## Snapshot

| | Gemini 3.5 Flash | GPT-5.4 |
|---|---|---|
| Provider | Google | OpenAI |
| Noometry Index | 54.1 | 59.5 |
| Rank | 29 | 14 |
| Context | 1.05M | 1.05M |
| Input $/M | $1.50 | $2.50 |
| Output $/M | $9 | $15 |
| Weights | Proprietary | Proprietary |

## Coding

- Gemini 3.5 Flash: 49.1 (#46)
- GPT-5.4: 52.8 (#28)

| Benchmark | Gemini 3.5 Flash | GPT-5.4 |
|---|---|---|
| SWE-bench Verified | 79.3% | 76.9% |
| DeepSWE | 37.4% | 51.8% |
| LMArena WebDev | 1499 | 1465 |
| SciCode | 53.1% | 56.6% |
| WeirdML | 62.6% | 77.7% |
| LMArena Coding | 1492 | 1497 |
| ALE-Bench | 911.02 | 1,607 |
| GSO | — | 31.4% |
| MirrorCode | — | 15.6% |
| AlgoTune | — | 1.85 |

## Agentic & Tool Use

- Gemini 3.5 Flash: 24.8 (#107)
- GPT-5.4: 46.5 (#12)

| Benchmark | Gemini 3.5 Flash | GPT-5.4 |
|---|---|---|
| APEX-Agents | 27.5% | 52.4% |
| GBAEval | 6.7% | 45.1% |
| Vending-Bench 2 | 5,396 | 6,144 |
| Terminal-Bench | — | 81.8% |
| τ²-bench Banking | — | 39.4% |
| DeepResearch Bench | — | 35.1% |
| PostTrainBench | — | 19% |
| GDP.pdf | 14% | — |
| LMArena Search | — | 1197 |
| METR Time Horizons | — | 74.3% |

## Reasoning

- Gemini 3.5 Flash: 62.7 (#16)
- GPT-5.4: 61.7 (#17)

| Benchmark | Gemini 3.5 Flash | GPT-5.4 |
|---|---|---|
| ARC-AGI-2 | 72.1% | 74% |
| NYT Connections (extended) | 92.6% | 91.3% |
| ARC-AGI-1 | 92.5% | 93.7% |
| CritPt | 13.1% | 23.4% |
| Chess Puzzles | 50% | 44% |
| EnigmaEval | 25.4% | 16% |
| EBR-Bench | 4.8% | 25.4% |
| LMArena Hard Prompts | 1488 | 1485 |
| Mystery Game Puzzles | 32% | 37% |
| DTBench | 94.7% | 94.4% |
| LMCA | 47.1% | 52% |
| Epoch Capabilities Index | 154.55 | 156.92 |
| ForecastBench | 59 | 59.5 |
| SimpleBench | 76.7% | — |
| Kagi LLM Benchmark | — | 63.8% |
| Thematic Generalization | — | 80% |
| Surface Evolver Bench | 58.1% | — |

## Math

- Gemini 3.5 Flash: 60.5 (#30)
- GPT-5.4: 73.3 (#14)

| Benchmark | Gemini 3.5 Flash | GPT-5.4 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 62.8% | 78.6% |
| FrontierMath Tier 4 | 26.8% | 49% |
| MathArena Final-Answer Competitions | 76.3% | 83.1% |
| OTIS Mock AIME 2024-2025 | 95.6% | 97.8% |
| ProofBench | 31% | 56% |
| LMArena Math | 1502 | 1488 |
| FrontierMath (Feb 2025 set) | 39% | 47.6% |
| FrontierMath Tier 4 (v1) | 14.6% | 27.1% |

## Knowledge

- Gemini 3.5 Flash: 66.3 (#10)
- GPT-5.4: 65.3 (#13)

| Benchmark | Gemini 3.5 Flash | GPT-5.4 |
|---|---|---|
| GPQA Diamond | 92.8% | 93.3% |
| SimpleQA Verified | 66.2% | 45.1% |
| LMArena Expert | 1493 | 1507 |
| Humanity's Last Exam | — | 36.2% |
| Vectara Hallucination Rate | — | 7% |

## Multimodal

- Gemini 3.5 Flash: 45.7 (#13)
- GPT-5.4: 43.8 (#19)

| Benchmark | Gemini 3.5 Flash | GPT-5.4 |
|---|---|---|
| LMArena Vision | 1311 | 1302 |
| Blueprint-Bench 2 | 33.6% | 27.1% |
| LMArena Document | 1463 | 1471 |
| Furniture Assembly | — | 37.5% |

## Multilingual

- Gemini 3.5 Flash: 57.1 (#12)
- GPT-5.4: 56.4 (#22)

| Benchmark | Gemini 3.5 Flash | GPT-5.4 |
|---|---|---|
| LMArena Non-English | 1475 | 1465 |
| LMArena Chinese | 1525 | 1520 |
| LMArena French | 1490 | 1493 |
| LMArena German | 1500 | 1472 |
| LMArena Japanese | 1481 | 1484 |
| LMArena Korean | 1453 | 1449 |
| LMArena Russian | 1491 | 1480 |
| LMArena Spanish | 1478 | 1453 |

## Instruction Following

- Gemini 3.5 Flash: 77.2 (#29)
- GPT-5.4: 77.3 (#25)

| Benchmark | Gemini 3.5 Flash | GPT-5.4 |
|---|---|---|
| LMArena Instruction Following | 1466 | 1469 |

## Long Context

- Gemini 3.5 Flash: 45.3 (#39)
- GPT-5.4: 50.3 (#8)

| Benchmark | Gemini 3.5 Flash | GPT-5.4 |
|---|---|---|
| LMArena Longer Query | 1481 | 1474 |
| CL-bench | — | 27.9% |
| CL-bench Life | — | 21.7% |

## Writing & Preference

- Gemini 3.5 Flash: 65.5 (#46)
- GPT-5.4: 72.1 (#18)

| Benchmark | Gemini 3.5 Flash | GPT-5.4 |
|---|---|---|
| LMArena Text | 1481 | 1469 |
| LMArena Creative Writing | 1467 | 1439 |
| EQ-Bench 4 | 1087 | 1272 |
| LMArena Multi-Turn | 1482 | 1482 |
| EQ-Bench Creative Writing | — | 1840 |

## FAQ

### Is Gemini 3.5 Flash better than GPT-5.4?

GPT-5.4 is the stronger model overall, scoring 59.5 to 54.1 on the Noometry Index. Gemini 3.5 Flash costs 1.7× less per token, which makes it the better buy when GPT-5.4's lead doesn't matter for your workload.

### Which is cheaper, Gemini 3.5 Flash or GPT-5.4?

Gemini 3.5 Flash is cheaper. It lists at $1.50 per million input tokens and $9 per million output tokens; GPT-5.4 lists at $2.50 and $15.

### Is Gemini 3.5 Flash or GPT-5.4 better for coding?

GPT-5.4 scores higher on coding benchmarks: 52.8 versus 49.1 in the Noometry coding category.

### Which has the bigger context window?

GPT-5.4 does, with 1.05M tokens against 1.05M.

### How many benchmarks do Gemini 3.5 Flash and GPT-5.4 share?

51 benchmarks have published results for both models. Gemini 3.5 Flash has 54 scored results on Noometry and GPT-5.4 has 68.