# Gemini 2.5 Flash vs GPT-5.4

> GPT-5.4 is the stronger model overall, scoring 59.5 to 39.0 on the Noometry Index. Gemini 2.5 Flash costs 6.6× 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-2-5-flash-vs-gpt-5-4
- Last updated: 2026-10-10
- Shared benchmarks: 37

## Summary

- They share 37 benchmarks with published results for both. Gemini 2.5 Flash scores higher in 0 categories and GPT-5.4 in 10 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5.4 leads 61.7 to 18.2.
- The biggest single-benchmark swing is ARC-AGI-2: 2.5% for Gemini 2.5 Flash and 74% for GPT-5.4.
- Gemini 2.5 Flash is cheaper at $0.30 / $2.50 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 2.5 Flash | GPT-5.4 |
|---|---|---|
| Provider | Google | OpenAI |
| Noometry Index | 39.0 | 59.5 |
| Rank | 162 | 14 |
| Context | 1.05M | 1.05M |
| Input $/M | $0.30 | $2.50 |
| Output $/M | $2.50 | $15 |
| Weights | Proprietary | Proprietary |

## Coding

- Gemini 2.5 Flash: 35.5 (#210)
- GPT-5.4: 52.8 (#28)

| Benchmark | Gemini 2.5 Flash | GPT-5.4 |
|---|---|---|
| WeirdML | 41.9% | 77.7% |
| LMArena Coding | 1424 | 1497 |
| ALE-Bench | 661.88 | 1,607 |
| SWE-bench Verified | — | 76.9% |
| DeepSWE | — | 51.8% |
| SWE-bench Verified (bash only) | 28.7% | — |
| Aider Polyglot | 55.1% | — |
| LMArena WebDev | — | 1465 |
| SciCode | — | 56.6% |
| GSO | — | 31.4% |
| MirrorCode | — | 15.6% |
| AlgoTune | — | 1.85 |

## Agentic & Tool Use

- Gemini 2.5 Flash: 30.6 (#67)
- GPT-5.4: 46.5 (#12)

| Benchmark | Gemini 2.5 Flash | GPT-5.4 |
|---|---|---|
| Terminal-Bench | 17.1% | 81.8% |
| Vending-Bench 2 | 548.84 | 6,144 |
| APEX-Agents | — | 52.4% |
| Berkeley Function Calling Leaderboard | 56.2% | — |
| TheAgentCompany | 41.1% | — |
| τ²-bench Banking | — | 39.4% |
| DeepResearch Bench | — | 35.1% |
| PostTrainBench | — | 19% |
| BALROG | 33.5% | — |
| GBAEval | — | 45.1% |
| LMArena Search | — | 1197 |
| METR Time Horizons | — | 74.3% |

## Reasoning

- Gemini 2.5 Flash: 18.2 (#271)
- GPT-5.4: 61.7 (#17)

| Benchmark | Gemini 2.5 Flash | GPT-5.4 |
|---|---|---|
| ARC-AGI-2 | 2.5% | 74% |
| Kagi LLM Benchmark | 56.8% | 63.8% |
| ARC-AGI-1 | 33.3% | 93.7% |
| CritPt | 1.1% | 23.4% |
| EnigmaEval | 2.7% | 16% |
| LMArena Hard Prompts | 1422 | 1485 |
| DTBench | 76.5% | 94.4% |
| LMCA | 27.5% | 52% |
| Epoch Capabilities Index | 143.04 | 156.92 |
| ForecastBench | 60.6 | 59.5 |
| SimpleBench | 41.2% | — |
| NYT Connections (extended) | — | 91.3% |
| Chess Puzzles | — | 44% |
| Thematic Generalization | — | 80% |
| EBR-Bench | — | 25.4% |
| Mystery Game Puzzles | — | 37% |

## Math

- Gemini 2.5 Flash: 38.1 (#87)
- GPT-5.4: 73.3 (#14)

| Benchmark | Gemini 2.5 Flash | GPT-5.4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 73.1% | 97.8% |
| LMArena Math | 1415 | 1488 |
| FrontierMath (Feb 2025 set) | 4.8% | 47.6% |
| FrontierMath Tier 4 (v1) | 4.2% | 27.1% |
| FrontierMath (Tiers 1-3) | — | 78.6% |
| FrontierMath Tier 4 | — | 49% |
| MathArena Final-Answer Competitions | — | 83.1% |
| ProofBench | — | 56% |
| Omni-MATH | 38.5% | — |

## Knowledge

- Gemini 2.5 Flash: 36.1 (#162)
- GPT-5.4: 65.3 (#13)

| Benchmark | Gemini 2.5 Flash | GPT-5.4 |
|---|---|---|
| Humanity's Last Exam | 12.1% | 36.2% |
| Vectara Hallucination Rate | 7.8% | 7% |
| LMArena Expert | 1426 | 1507 |
| GPQA Diamond | — | 93.3% |
| SimpleQA Verified | — | 45.1% |
| MMLU-Pro | 63.9% | — |
| Confabulations | 16.8% | — |
| GPQA (HELM) | 39% | — |

## Multimodal

- Gemini 2.5 Flash: 41.2 (#31)
- GPT-5.4: 43.8 (#19)

| Benchmark | Gemini 2.5 Flash | GPT-5.4 |
|---|---|---|
| LMArena Vision | 1254 | 1302 |
| GeoBench | 76% | — |
| VPCT | 46.2% | — |
| Blueprint-Bench 2 | — | 27.1% |
| Furniture Assembly | — | 37.5% |
| LMArena Document | — | 1471 |
| SpatialViz-Bench | 36.9% | — |

## Multilingual

- Gemini 2.5 Flash: 52.5 (#83)
- GPT-5.4: 56.4 (#22)

| Benchmark | Gemini 2.5 Flash | GPT-5.4 |
|---|---|---|
| LMArena Non-English | 1410 | 1465 |
| LMArena Chinese | 1450 | 1520 |
| LMArena French | 1433 | 1493 |
| LMArena German | 1419 | 1472 |
| LMArena Japanese | 1403 | 1484 |
| LMArena Korean | 1385 | 1449 |
| LMArena Russian | 1415 | 1480 |
| LMArena Spanish | 1421 | 1453 |

## Instruction Following

- Gemini 2.5 Flash: 75.9 (#50)
- GPT-5.4: 77.3 (#25)

| Benchmark | Gemini 2.5 Flash | GPT-5.4 |
|---|---|---|
| LMArena Instruction Following | 1405 | 1469 |
| IFEval | 89.8% | — |

## Long Context

- Gemini 2.5 Flash: 47.5 (#16)
- GPT-5.4: 50.3 (#8)

| Benchmark | Gemini 2.5 Flash | GPT-5.4 |
|---|---|---|
| LMArena Longer Query | 1419 | 1474 |
| Fiction.LiveBench | 77.8% | — |
| CL-bench | — | 27.9% |
| CL-bench Life | — | 21.7% |

## Writing & Preference

- Gemini 2.5 Flash: 54.0 (#153)
- GPT-5.4: 72.1 (#18)

| Benchmark | Gemini 2.5 Flash | GPT-5.4 |
|---|---|---|
| LMArena Text | 1417 | 1469 |
| LMArena Creative Writing | 1400 | 1439 |
| EQ-Bench Creative Writing | 1137 | 1840 |
| LMArena Multi-Turn | 1408 | 1482 |
| Short-Story Creative Writing | 76.5% | — |
| WildBench | 81.7% | — |
| EQ-Bench 4 | — | 1272 |

## FAQ

### Is Gemini 2.5 Flash better than GPT-5.4?

GPT-5.4 is the stronger model overall, scoring 59.5 to 39.0 on the Noometry Index. Gemini 2.5 Flash costs 6.6× 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 2.5 Flash or GPT-5.4?

Gemini 2.5 Flash is cheaper. It lists at $0.30 per million input tokens and $2.50 per million output tokens; GPT-5.4 lists at $2.50 and $15.

### Is Gemini 2.5 Flash or GPT-5.4 better for coding?

GPT-5.4 scores higher on coding benchmarks: 52.8 versus 35.5 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 2.5 Flash and GPT-5.4 share?

37 benchmarks have published results for both models. Gemini 2.5 Flash has 54 scored results on Noometry and GPT-5.4 has 68.