AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

UEval: A Benchmark for Unified Multimodal Generation

Bo Li, Yida Yin, Wenhao Chai, Xingyu Fu, Zhuang Liu

ECCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Multimodal LLM benchmarks · 14 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
append GPT-5-Thinking reasoning trace to the prompt (ablation)GPT-5-Instant · UEval image generationimage qualityImproved
append GPT-5-Thinking reasoning trace to the prompt (ablation)Gemini-2.5-Flash · UEval image generationimage qualityImproved
append GPT-5-Thinking reasoning trace to the prompt (ablation)BAGEL · UEvalgeneration qualityNo clear effect
use reasoning model GPT-5-Thinking instead of GPT-5-InstantGPT-5 · UEval avgrubric score: 65.2 -> 66.4Improved
use reasoning model GPT-5-Thinking instead of GPT-5-InstantGPT-5 · UEval Paperrubric score: 55.1 -> 51.9Hurt
use proprietary GPT-5-Thinking instead of best open-source Emu3.5unified models · UEval avgrubric score: 49.1 -> 66.4Improved
generate text in unified models instead of imagesGPT-5-Thinking · UEvalrubric score: 49.1 -> 83.8Improved
open-source unified model text versus text-only modelEmu3.5 · UEval texttext score: 64.6 -> 81.2Hurt
multi-step planning tasks versus knowledge tasksGPT-5-Thinking · UEvalrubric score: 78.0 -> 67.8Hurt
use Seed1.6-Vision as judge instead of Gemini-2.5-Pro (ablation)GPT-5-Thinking responses · UEval Artrubric score: 67.8 -> 75.2Mixed
use GLM-4.1V-Thinking as judge instead of Gemini-2.5-Pro (ablation)GPT-5-Thinking responses · UEval Techrubric score: 57.0 -> 74.3Mixed
use Qwen3-VL-235B-Thinking as judge instead of Gemini-2.5-Pro (ablation)GPT-5-Thinking responses · UEval Spacerubric score: 84.0 -> 81.8No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy