GSM8K-V: leaderboard
Each model solves all 1,319 GSM8K test problems redrawn as two to eleven images, seeing only the images and a format prompt, and the score is the share of correct final numbers. Unlike GSM8K and MathVista, no problem text is given, so the quantities and steps must be read from the scenes.
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Around February 2027. 21 models tracked. Updated 17 Nov 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 59.06 | #120 | |
| 2 | Kimi K2.6 | 47.34 | #191 | |
| 3 | GLM-4.6V | 37.78 | #710 | |
| 4 | Llama 4 Maverick Instruct | 31.69 | #1151 | |
| 5 | GPT-5 | Minimal | 30.1 | #177 |
| 6 | Qwen 3 VL 30B A3B (Thinking) | 28.58 | #728 | |
| 7 | GLM-4.5V | 28.28 | #758 | |
| 8 | Qwen 3 VL 30B A3B Instruct | 27.9 | #894 | |
| 9 | InternVL3.5-241B-A28B | 27.9 | #720 | |
| 10 | Ovis2.5-9B | 27.45 | #1237 | |
| 11 | Llama 4 Scout Instruct | 26.69 | #1575 | |
| 12 | Qwen 3 VL 8B (Thinking) | 26.46 | #914 | |
| 13 | Qwen 3 VL 8B Instruct | 26 | #1025 | |
| 14 | Step-3 | 25.85 | #739 | |
| 15 | InternVL3.5-30B-A3B | 22.82 | #1265 |
No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.
Interactive version: theaggregate.ai/benchmark?slug=gsm8k-v · How It Works · Data refreshed daily, snapshot 2026-11-17.