# Gemini 3.8 Flash vs GPT-5 Mini

> Gemini 3.8 Flash is the stronger model overall, scoring 61.8 to 41.8 on the Noometry Index. GPT-5 Mini costs 2.2× less per token, which makes it the better buy when Gemini 3.8 Flash's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/gemini-3-8-flash-vs-gpt-5-mini
- Last updated: 2026-10-11
- Shared benchmarks: 38

## Summary

- They share 38 benchmarks with published results for both. Gemini 3.8 Flash scores higher in 10 categories and GPT-5 Mini in 0 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Gemini 3.8 Flash leads 76.9 to 23.9.
- The biggest single-benchmark swing is ARC-AGI-2: 89.2% for Gemini 3.8 Flash and 4.4% for GPT-5 Mini.
- GPT-5 Mini is cheaper at $0.25 / $2 per million input/output tokens, against $0.75 / $3.75 for Gemini 3.8 Flash.
- Gemini 3.8 Flash accepts more context: 1.05M tokens versus 400K.

## Snapshot

| | Gemini 3.8 Flash | GPT-5 Mini |
|---|---|---|
| Provider | Google | OpenAI |
| Noometry Index | 61.8 | 41.8 |
| Rank | 11 | 128 |
| Context | 1.05M | 400K |
| Input $/M | $0.75 | $0.25 |
| Output $/M | $3.75 | $2 |
| Weights | Proprietary | Proprietary |

## Coding

- Gemini 3.8 Flash: 59.2 (#15)
- GPT-5 Mini: 40.1 (#146)

| Benchmark | Gemini 3.8 Flash | GPT-5 Mini |
|---|---|---|
| SciCode | 56.6% | 39.2% |
| WeirdML | 84.8% | 52.7% |
| LMArena Coding | 1510 | 1406 |
| ALE-Bench | 1,270 | 799.77 |
| SWE-bench Verified | — | 64.7% |
| DeepSWE | 73.8% | — |
| FrontierCode | 41.2% | — |
| SWE-bench Verified (bash only) | — | 59.8% |
| CursorBench | 39.6% | — |
| LMArena WebDev | 1584 | — |
| SWE-bench Multilingual | — | 39.7% |
| FrontierSWE | 19.6% | — |
| AlgoTune | — | 1.38 |

## Agentic & Tool Use

- Gemini 3.8 Flash: 41.8 (#21)
- GPT-5 Mini: 31.1 (#70)

| Benchmark | Gemini 3.8 Flash | GPT-5 Mini |
|---|---|---|
| Vending-Bench 2 | 5,094 | -31.18 |
| Terminal-Bench | — | 34.8% |
| APEX-Agents | 64.3% | — |
| Berkeley Function Calling Leaderboard | — | 55.5% |
| Remote Labor Index | 5.8% | — |
| GDP.pdf | 23.4% | — |

## Reasoning

- Gemini 3.8 Flash: 76.9 (#5)
- GPT-5 Mini: 23.9 (#168)

| Benchmark | Gemini 3.8 Flash | GPT-5 Mini |
|---|---|---|
| ARC-AGI-2 | 89.2% | 4.4% |
| ARC-AGI-1 | 98.5% | 54.3% |
| CritPt | 18.3% | 0% |
| Chess Puzzles | 61% | 30% |
| LMArena Hard Prompts | 1508 | 1380 |
| Mystery Game Puzzles | 47% | 10% |
| DTBench | 95.7% | 80.5% |
| LMCA | 52.9% | 34.2% |
| Epoch Capabilities Index | 156.71 | 145.52 |
| Kagi LLM Benchmark | — | 70.3% |
| NYT Connections (extended) | 97.4% | — |
| EnigmaEval | — | 8.2% |
| Surface Evolver Bench | 76.9% | — |
| ForecastBench | — | 61 |

## Math

- Gemini 3.8 Flash: 65.3 (#28)
- GPT-5 Mini: 46.7 (#69)

| Benchmark | Gemini 3.8 Flash | GPT-5 Mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 68.4% | 46.7% |
| FrontierMath Tier 4 | 22% | 12.2% |
| OTIS Mock AIME 2024-2025 | 98.9% | 86.7% |
| ProofBench | 48% | 9% |
| LMArena Math | 1528 | 1378 |
| Omni-MATH | — | 72.2% |
| MATH Level 5 | — | 97.8% |
| FrontierMath (Feb 2025 set) | — | 27.2% |
| FrontierMath Tier 4 (v1) | — | 6.3% |

## Knowledge

- Gemini 3.8 Flash: 74.8 (#2)
- GPT-5 Mini: 45.6 (#86)

| Benchmark | Gemini 3.8 Flash | GPT-5 Mini |
|---|---|---|
| GPQA Diamond | 95.4% | 75% |
| Humanity's Last Exam | 44.5% | 19.4% |
| SimpleQA Verified | 69.7% | 21.6% |
| LMArena Expert | 1524 | 1379 |
| MMLU-Pro | — | 83.5% |
| Confabulations | — | 13.3% |
| Vectara Hallucination Rate | — | 12.9% |
| GPQA (HELM) | — | 75.6% |

## Multimodal

- Gemini 3.8 Flash: 40.7 (#45)
- GPT-5 Mini: 35.6 (#85)

| Benchmark | Gemini 3.8 Flash | GPT-5 Mini |
|---|---|---|
| LMArena Vision | 1314 | 1202 |
| VPCT | — | 40.2% |
| Blueprint-Bench 2 | 38.6% | — |
| Furniture Assembly | 31.7% | — |

## Multilingual

- Gemini 3.8 Flash: 58.0 (#5)
- GPT-5 Mini: 48.9 (#137)

| Benchmark | Gemini 3.8 Flash | GPT-5 Mini |
|---|---|---|
| LMArena Non-English | 1491 | 1363 |
| LMArena Chinese | 1554 | 1385 |
| LMArena French | 1498 | 1386 |
| LMArena German | 1493 | 1366 |
| LMArena Japanese | 1502 | 1341 |
| LMArena Korean | 1459 | 1308 |
| LMArena Russian | 1515 | 1362 |
| LMArena Spanish | 1485 | 1355 |

## Instruction Following

- Gemini 3.8 Flash: 78.0 (#13)
- GPT-5 Mini: 76.2 (#46)

| Benchmark | Gemini 3.8 Flash | GPT-5 Mini |
|---|---|---|
| LMArena Instruction Following | 1490 | 1357 |
| IFEval | — | 92.7% |

## Long Context

- Gemini 3.8 Flash: 46.3 (#24)
- GPT-5 Mini: 41.9 (#132)

| Benchmark | Gemini 3.8 Flash | GPT-5 Mini |
|---|---|---|
| LMArena Longer Query | 1508 | 1355 |
| Fiction.LiveBench | — | 69.4% |

## Writing & Preference

- Gemini 3.8 Flash: 72.2 (#15)
- GPT-5 Mini: 55.2 (#148)

| Benchmark | Gemini 3.8 Flash | GPT-5 Mini |
|---|---|---|
| LMArena Text | 1499 | 1373 |
| LMArena Creative Writing | 1492 | 1325 |
| EQ-Bench Creative Writing | 1748 | 1313 |
| LMArena Multi-Turn | 1501 | 1363 |
| Short-Story Creative Writing | — | 83.1% |
| WildBench | — | 85.5% |

## FAQ

### Is Gemini 3.8 Flash better than GPT-5 Mini?

Gemini 3.8 Flash is the stronger model overall, scoring 61.8 to 41.8 on the Noometry Index. GPT-5 Mini costs 2.2× less per token, which makes it the better buy when Gemini 3.8 Flash's lead doesn't matter for your workload.

### Which is cheaper, Gemini 3.8 Flash or GPT-5 Mini?

GPT-5 Mini is cheaper. It lists at $0.25 per million input tokens and $2 per million output tokens; Gemini 3.8 Flash lists at $0.75 and $3.75.

### Is Gemini 3.8 Flash or GPT-5 Mini better for coding?

Gemini 3.8 Flash scores higher on coding benchmarks: 59.2 versus 40.1 in the Noometry coding category.

### Which has the bigger context window?

Gemini 3.8 Flash does, with 1.05M tokens against 400K.

### How many benchmarks do Gemini 3.8 Flash and GPT-5 Mini share?

38 benchmarks have published results for both models. Gemini 3.8 Flash has 50 scored results on Noometry and GPT-5 Mini has 60.