AI Paper Atlas

Research map / Benchmarks / ScanRefer

ScanRefer leaderboard

23 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on ScanRefer. Best Acc@0.25: 63.9 (LLaVA-3D (two-stage), ICCV 2025).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

ScanRefer · Acc@0.25

22 papers, each with its best reported number for its own method; higher is better.

#Acc@0.25 ↑MethodPaperTopic
163.9LLaVA-3D (two-stage)LLaVA-3D: A Simple yet Effective Pathway to Empowering LMMs with 3D Capabilities
ICCV 2025 · Table 5
Multimodal LLMs
263.3MVRD
LLaVA-Video-7B
Multi-view Relational Distillation for Spatial Reasoning with Vision-Language Models
NeurIPS 2026 · Table 7
Multimodal LLMs
363.2Vid-LLM (Ours)Vid-LLM: A Compact Video-based 3D Multimodal LLM with Reconstruction–Reasoning Synergy
ICLR 2026 · Table 3
Multimodal LLMs
463.2VEGA-3D
Video-3D LLM
Generation Models Know Space: Unleashing Implicit 3D Priors for Scene Understanding
ECCV 2026 · Table 1
Video generation
562.93DRS
LLaVA-Next-Video
MLLMs Need 3D-Aware Representation Supervision for Scene Understanding
NeurIPS 2025 · Table: Performance comparison
Multimodal LLMs
662.43DGraphLLM
LLAMA3-8B-Instruct
3DGraphLLM: Combining Semantic Graphs and Large Language Models for 3D Scene Understanding
ICCV 2025 · Table 1
3D generation
760.93D-IDE (Ours, RGB-only)
Qwen2-7B (LLaVA-Next-Video style)
3D-IDE: 3D Implicit Depth Emergent
CVPR 2026 · Table 2
3D generation
860.8Robin3D
Vicuna-7B-v1.5
Robin3D: Improving 3D Large Language Model via Robust Instruction Tuning
ICCV 2025 · Table 1
Multimodal LLMs
960.5APEIRIA with modular inference enhancement
8B MLLM backbone
Distilling Neuro-Symbolic Programs into 3D Multi-modal LLMs
ICML 2026 · Table 1
Agents, code and math
1059.8Chat-Scene++ (3D+2D)
Vicuna-7B
Chat-Scene++: Exploiting Context-Rich Object Identification for 3D LLM
TPAMI 2026 · Table 4
3D generation
12 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

ScanRefer · Acc@0.5

16 papers, each with its best reported number for its own method; higher is better.

6 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.