ScienceQA leaderboard
As of 6 Nov 2026, the top score on ScienceQA is 91.3% by Phi 3.5 Vision (Microsoft), ahead of GPT-4o at 88.5%, across 23 model configurations with a published result (Epoch AI Benchmarking Hub, result dated 16 Aug 2024). Science questions with diagrams. ScienceQA is shown for reference only and does not count toward the BenchLeader Index.
Key facts (as of 6 Nov 2026)
- Phi 3.5 Vision holds the top ScienceQA score at 91.3% (Epoch AI Benchmarking Hub, 16 Aug 2024).
- GPT-4o is second on ScienceQA at 88.5% (Epoch AI Benchmarking Hub, 13 May 2024).
- ScienceQA: Science questions with diagrams. A 2022 benchmark, saturated for current models; kept for history (Epoch AI Benchmarking Hub).
- 23 model configurations have a published ScienceQA result on BenchLeader (BenchLeader, 6 Nov 2026).
- ScienceQA does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).
Science questions with diagrams. A 2022 benchmark, saturated for current models; kept for history.
- Published by
- Epoch AI Benchmarking Hub
- Category
- Multimodal
- Index weight
- Reference only
- Models
- 23
- 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
Science questions with diagrams.
How it is scored
Accuracy, as compiled by Epoch AI from published results.
What to keep in mind
Saturated: frontier models score near the ceiling, so it separates only older or smaller models, and results are largely self-reported by developers.
- 1Phi 3.5 Vision91.3%
- 2GPT-4o88.5%
- 3Gemini 1.0 Pro Vision79.7%
- 4falcon-11B-vlm74.9%
- 5Blip2 Opt 2 7B74.2%
- 6InstructGPT 175B74.0%
- 7LLaVA-NeXT 8B (Llama 3)73.7%
- 8LLaVA-1.6 Vicuna 13B73.6%
- 9LLaVA-1.6 Mistral 7B72.8%
- 10MM1-7B-Chat72.6%
- 11Claude 3 Haiku72.0%
- 12LLaVA-1.6 Vicuna 7B70.6%
- 13InternVL-Chat-ViT-6B-Vicuna-13B70.1%
- 14MM1-3B-Chat69.4%
- 15Qwen-VL68.2%
| # | ||||
|---|---|---|---|---|
| 1 |  Phi 3.5 Vision | 91.3% | – | 2024-08-16 |
| 2 |  GPT-4o | 88.5% | 43.3 | 2024-05-13 |
| 3 |  Gemini 1.0 Pro Vision | 79.7% | – | 2024-01-04 |
| 4 |  falcon-11B-vlm | 74.9% | – | 2024-05-21 |
| 5 |  Blip2 Opt 2 7B | 74.2% | – | 2023-02-06 |
| 6 | InstructGPT 175B | 74.0% | – | 2022-01-27 |
| 7 |  LLaVA-NeXT 8B (Llama 3) | 73.7% | – | 2024-04-20 |
| 8 |  LLaVA-1.6 Vicuna 13B | 73.6% | – | 2024-01-31 |
| 9 | LLaVA-1.6 Mistral 7B | 72.8% | – | 2024-01-31 |
| 10 |  MM1-7B-Chat | 72.6% | – | 2024-03-14 |
| 11 |  Claude 3 Haiku | 72.0% | 36.3 | 2024-03-07 |
| 12 | LLaVA-1.6 Vicuna 7B | 70.6% | – | 2024-01-31 |
| 13 |  InternVL-Chat-ViT-6B-Vicuna-13B | 70.1% | – | 2024-08-16 |
| 14 | MM1-3B-Chat | 69.4% | – | 2024-03-14 |
| 15 |  Qwen-VL | 68.2% | – | 2023-08-20 |
| 16 | LLaVA-1.5 7B | 66.8% | – | 2023-10-05 |
| 17 | InternVL-Chat-ViT-6B-Vicuna-7B | 66.2% | – | 2023-12-25 |
| 18 |  InstructBLIP Vicuna 13B | 63.1% | – | 2023-12-25 |
| 19 | InstructBLIP Vicuna 7B | 60.5% | – | 2023-05-22 |
| 20 | Llama 2-13Bopen ↗ | 55.8% | – | 2023-07-18 |
| 21 | LLaMA-13B | 43.3% | – | 2023-02-24 |
| 22 | Llama 2-7Bopen ↗ | 43.1% | – | 2023-07-18 |
| 23 | LLaMA-7B | 36.2% | – | 2023-02-24 |
More multimodal benchmarks · All LLM benchmarks
Related benchmarks
Cite as: BenchLeader, “ScienceQA leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/scienceqa.
ScienceQA: questions
- What does ScienceQA measure?
- Science questions with diagrams. Scores are reported in percent of tasks solved; higher is better.
- What is the top score on ScienceQA?
- As of 6 Nov 2026, the highest score on ScienceQA is 91.3% by Phi 3.5 Vision (result dated 16 Aug 2024), ahead of GPT-4o at 88.5%. Scores come from Epoch AI Benchmarking Hub and are re-read daily.
- Which AI model leads ScienceQA?
- Phi 3.5 Vision leads ScienceQA with 91.3% as of 6 Nov 2026 (result dated 16 Aug 2024), ahead of GPT-4o at 88.5%.
- How many models have ScienceQA results?
- 23 model configurations have a ScienceQA result on BenchLeader, all taken from Epoch AI Benchmarking Hub.
- Who runs ScienceQA and how often is it updated?
- ScienceQA is published by Epoch AI Benchmarking Hub. BenchLeader re-reads the published results once a day and records the date each result was published.
- Does ScienceQA count toward the BenchLeader Index?
- No. ScienceQA is shown for reference but left out of the composite index.