Visual Quality Evaluation for Presentation Refinement on CoEdit Eval
0Base Win Countopenai/gpt-4o
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| openai/gpt-4oN=10, Judge=GPT-4o2026.05 | 0 | — | — | 4 | — | 4 | — | 3.35 | — | |
| Pro/zai-org/GLM-4VN=9, Judge=GPT-4o2026.05 | 0 | — | — | 4.11 | — | 4.11 | — | 3.44 | — | |
| Qwen/Qwen2-VL-7BN=8, Judge=GPT-4o2026.05 | 0 | — | — | 4.06 | — | 4.06 | — | 3.56 | — | |
| google/gemini-1.5-proN=6, Judge=GPT-4o2026.05 | 0 | — | — | 4 | — | 4 | — | 3.42 | — | |
| ECHOBase Model=openai/gpt-4o, N=10, Judge=GPT-4o2026.05 | — | 9 | 1 | — | 4.6 | — | 4.4 | — | 4.75 | |
| ECHOBase Model=Pro/zai-org/GLM-4V, N=9, Judge=GPT-4o2026.05 | — | 7 | 2 | — | 4.5 | — | 4.39 | — | 4.67 | |
| ECHOBase Model=Qwen/Qwen2-VL-7B, N=8, Judge=GPT-4o2026.05 | — | 7 | 1 | — | 4.63 | — | 4.44 | — | 4.69 | |
| ECHOBase Model=google/gemini-1.5-pro, N=6, Judge=GPT-4o2026.05 | — | 5 | 1 | — | 4.5 | — | 4.33 | — | 4.67 |