AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

SpecEyes: Accelerating Agentic Multimodal LLM via Speculative Planning and Perception

Haoyu Huang, Jinfa Huang, Zhongwei Wan, Xiawu Zheng, Rongrong Ji, Jiebo Luo

ECCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Chain-of-thought multimodal reasoning

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
route queries with min-separability gating to a small tool-free model before the agentic modelDeepEyes · Average (V*, HR-Bench, POPE)accuracy / speedup: 81.39 -> 84.26 at 1.73xImproved
route queries with min-separability gatingThyme · Average (V*, HR-Bench, POPE)accuracy / speedup: 82.29 -> 83.99 at 1.42xImproved
apply SpecEyes to high-resolution inputsSpecEyes (min) · HR-Bench 8Kspeedup: 1.08x (DeepEyes), 0.95x (Thyme)No clear effect
use min aggregation of token separability instead of mean (ablation)SpecEyes (DeepEyes) · Average (V*, HR-Bench, POPE)accuracy: 80.53 -> 84.26Improved
use separability score instead of softmax geometric-mean confidence (ablation)SpecEyes (DeepEyes) · Average (V*, HR-Bench, POPE)accuracy: 82.31 -> 84.26Improved
use min instead of bottom-r aggregation (ablation)SpecEyes (Thyme) · Average (V*, HR-Bench, POPE)accuracy: 81.95 -> 83.99Improved
disable tools in the large backbone instead of routing (ablation)DeepEyes · POPE Adversarialaccuracy: 78.43 -> 46.90Hurt
replace Qwen3-VL-2B draft with larger Qwen3-VL-8B (ablation)SpecEyes (min), DeepEyes · Average (V*, HR-Bench, POPE)accuracy / speedup: V* Attr 92.17 vs 90.43; speedup 1.67x vs 1.73xMixed
increase Top-K in separability computation (ablation)SpecEyes (min) · V* and HR-Benchaccuracy / speedupMixed
lower the gating threshold (ablation)SpecEyes (min) · V*, HR-Bench, POPEaccuracy / speedupMixed
increase serving batch size (ablation)SpecEyes · V*, HR-Bench, POPEspeedupImproved
use token-level speculative reasoning (SpecReason) instead of agentic-level speculationDeepEyes · Average (V*, HR-Bench, POPE)speedup: 0.43x averageHurt

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