GeoBench leaderboard
As of 6 Nov 2026, the top score on GeoBench is 88.0% by Gemini 3 Flash (Google), ahead of Gemini 2.5 Pro at 86.0%, across 27 model configurations with a published result (GeoBench, result dated 17 Dec 2025). Given a street-level photo, the model must name the country (and finer levels) where it was taken. GeoBench is shown for reference only and does not count toward the BenchLeader Index.
Key facts (as of 6 Nov 2026)
- Gemini 3 Flash holds the top GeoBench score at 88.0% (GeoBench, 17 Dec 2025).
- Gemini 2.5 Pro is second on GeoBench at 86.0% (GeoBench, 6 May 2025).
- GeoBench: Identifying where a photo was taken (GeoBench).
- 27 model configurations have a published GeoBench result on BenchLeader (BenchLeader, 6 Nov 2026).
- GeoBench does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).
Identifying where a photo was taken.
- Published by
- GeoBenchdata via Epoch AI Benchmarking Hub
- Category
- Multimodal
- Index weight
- Reference only
- Models
- 27
- Data as of
- 6 Nov 2026
CC BY 4.0 — Epoch AI, ‘AI Benchmarking Hub’, epoch.ai/benchmarks. Mirrored boards credit their original publishers.
What the test looks like
Given a street-level photo, the model must name the country (and finer levels) where it was taken.
How it is scored
Percent of countries identified, as published by GeoBench.
What to keep in mind
Vision models only; a test of world knowledge through images.
- 1Gemini 3 Flash88.0%
- 2Gemini 2.5 Pro86.0%
- 3Gemini 3 Pro84.0%
- 4GPT-5 (medium)81.0%
- 5o1 (medium)80.0%
- 6Gemini 2.0 Flash77.0%
- 7Gemini 2.5 Flash76.0%
- 8Gemini 1.5 Flash 00276.0%
- 9Claude Opus 4.575.0%
- 10o3 (medium)74.0%
- 11GPT-4.172.0%
- 12GPT-4o71.0%
- 13Claude 3.7 Sonnet68.0%
- 14o4-mini (medium)64.0%
- 15o4-mini (high)64.0%
| # | ||||
|---|---|---|---|---|
| 1 |  Gemini 3 Flash | 88.0% | 56.9 | 2025-12-17 |
| 2 | Gemini 2.5 Pro | 86.0% | 54.5 | 2025-05-06 |
| 3 | Gemini 3 Pro | 84.0% | 60.2 | 2025-11-18 |
| 4 |  GPT-5medium | 81.0% | 58.1 | 2025-08-07 |
| 5 | o1medium | 80.0% | 48.1 | 2024-12-17 |
| 6 | Gemini 2.0 Flash | 77.0% | 43.5 | 2025-02-05 |
| 7 | Gemini 2.5 Flash | 76.0% | 52.7 | 2025-05-20 |
| 8 | Gemini 1.5 Flash 002 | 76.0% | 38.2 | 2024-09-24 |
| 9 |  Claude Opus 4.5 | 75.0% | 57.9 | 2025-11-24 |
| 10 | o3medium | 74.0% | 53.3 | 2025-04-16 |
| 11 | GPT-4.1 | 72.0% | 48.5 | 2025-04-14 |
| 12 | GPT-4o | 71.0% | 43.3 | 2024-11-20 |
| 13 | Claude 3.7 Sonnet | 68.0% | 49.9 | 2025-02-24 |
| 14 | o4-minimedium | 64.0% | 51.8 | 2025-04-16 |
| 15 | o4-minihigh | 64.0% | 50.3 | 2025-04-16 |
| 16 | GPT-4o mini | 64.0% | 35.4 | 2024-07-18 |
| 17 | Claude 3.5 Sonnet | 62.0% | 47.1 | 2024-10-22 |
| 18 |  Qwen2.5 VL 72B Instructopen ↗ | 62.0% | – | 2024-09-19 |
| 19 | o3high | 60.0% | 53.8 | 2025-04-16 |
| 20 |  Llama 4 Maverickopen ↗ | 52.0% | 42.1 | 2025-04-05 |
| 21 | Gemma 3 27Bopen ↗ | 52.0% | 39.9 | 2025-03-12 |
| 22 | Llama 3.2 90B Visionopen ↗ | 52.0% | 36.1 | 2024-09-24 |
| 23 | Claude Opus 4 | 49.0% | 52.6 | 2025-05-22 |
| 24 |  Grok 4 | 45.0% | 58.3 | 2025-07-09 |
| 25 | Claude Sonnet 4 | 37.0% | 50.5 | 2025-05-22 |
| 26 |  Pixtral 12Bopen | 37.0% | – | 2024-09-17 |
| 27 | Claude 3.5 Haiku | 34.0% | 38.8 | 2024-10-22 |
More multimodal benchmarks · All LLM benchmarks
Related benchmarks
Cite as: BenchLeader, “GeoBench leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/geobench.
GeoBench: questions
- What does GeoBench measure?
- Given a street-level photo, the model must name the country (and finer levels) where it was taken. Scores are reported in percent of tasks solved; higher is better.
- What is the top score on GeoBench?
- As of 6 Nov 2026, the highest score on GeoBench is 88.0% by Gemini 3 Flash (result dated 17 Dec 2025), ahead of Gemini 2.5 Pro at 86.0%. Scores come from GeoBench and are re-read daily.
- Which AI model leads GeoBench?
- Gemini 3 Flash leads GeoBench with 88.0% as of 6 Nov 2026 (result dated 17 Dec 2025), ahead of Gemini 2.5 Pro at 86.0%.
- How many models have GeoBench results?
- 27 model configurations have a GeoBench result on BenchLeader, all taken from GeoBench via Epoch AI Benchmarking Hub.
- Who runs GeoBench and how often is it updated?
- GeoBench is published by GeoBench. BenchLeader re-reads the published results once a day and records the date each result was published.
- Does GeoBench count toward the BenchLeader Index?
- No. GeoBench is shown for reference but left out of the composite index.