Multimodal Perception on HR-8K
85AccuracyGemini-3-Flash
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini-3-Flash2026.02 | 85 | |
| ZwZ-8Bnumber of parameters=8B2026.02 | 82 | |
| ZwZModel Size=8B2026.02 | 82 | |
| Qwen3-VL-235B-A22Bnumber of parameters=235B2026.02 | 81.62 | |
| TACO (Ours-7B)Model Category=Code-tool / visual-agent models2026.06 | 81.6 | |
| Gemini-2.5-ProModel Category=Closed-source models2026.06 | 81.5 | |
| ZwZ-4Bnumber of parameters=4B2026.02 | 79.5 | |
| Kimi-K2.5 (1T)number of parameters=1T2026.02 | 75.38 | |
| GLM-4.5V (108B)number of parameters=108B2026.02 | 74.88 | |
| Qwen3-VL-8Bnumber of parameters=8B2026.02 | 74.63 | |
| Qwen3-VL-8BModel Category=Open-source MLLMs (no visual tool), Instruct variant=true2026.06 | 74.6 | |
| PyVision-RL-7BModel Category=Code-tool / visual-agent models2026.06 | 74.3 | |
| GPT-5Model Category=Closed-source models2026.06 | 74.1 | |
| DeepEyes-v2-7BModel Category=Code-tool / visual-agent models2026.06 | 73.8 | |
| ZwZModel Size=7B2026.02 | 73.3 | |
| Mini-o3-7BModel Category=Code-tool / visual-agent models2026.06 | 73.3 | |
| ZwZ-7Bnumber of parameters=7B2026.02 | 73.25 | |
| Qwen3-VL-4Bnumber of parameters=4B2026.02 | 72.88 | |
| MiMo-VL-7B-RLnumber of parameters=7B2026.02 | 72.88 | |
| DeepEyes-7BModel Category=Code-tool / visual-agent models2026.06 | 72.6 | |
| Thyme-7BModel Category=Code-tool / visual-agent models2026.06 | 72 | |
| CodeV-RL-7BModel Category=Code-tool / visual-agent models2026.06 | 71.3 | |
| Image-JigsawModel Size=7B2026.02 | 71.1 | |
| MathCoder-VL-8BModel Category=Code-tool / visual-agent models2026.06 | 70.6 | |
| AGILEModel Size=7B2026.02 | 70.5 | |
| DiGModel Size=8B2026.02 | 70.4 | |
| Qwen2.5-VL-32BModel Category=Open-source MLLMs (no visual tool), Instruct variant=true2026.06 | 70.4 | |
| InternVL3-8BModel Category=Open-source MLLMs (no visual tool)2026.06 | 69.3 | |
| Qwen2.5-VL-7Bnumber of parameters=7B2026.02 | 67.88 | |
| Pixel-Reasoner-7BModel Category=Code-tool / visual-agent models2026.06 | 66.9 | |
| Qwen2.5-VL-7BModel Category=Open-source MLLMs (no visual tool), Instruct variant=true2026.06 | 65.3 | |
| GPT-5.12026.02 | 65.25 | |
| MiniCPM-V-4.5 (9B)number of parameters=9B2026.02 | 63.62 | |
| LLaVA-OneVision-7BModel Category=Open-source MLLMs (no visual tool)2026.06 | 59.8 | |
| GPT-4oModel Category=Closed-source models2026.06 | 59.6 |