AI Paper Atlas

Research map / Benchmarks / ScanQA

ScanQA leaderboard

64 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on ScanQA. Best CIDEr (val): 109.3 (SR-3D-8B, ICLR 2026).

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.

ScanQA · CIDEr (val)

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

28 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

ScanQA · CIDEr

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

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

ScanQA · EM@1 (val)

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

#EM@1 ↑MethodPaperTopic
130.6LLaVA-3DLLaVA-3D: A Simple yet Effective Pathway to Empowering LMMs with 3D Capabilities
ICCV 2025 · Table 1
Multimodal LLMs
230.4SpaR3D-MoESpaR3D-MoE: Adaptive 3D Spatial Reasoning from Sparse Views Meets Geometry-Inductive Mixture-of-Experts
ECCV 2026 · Table 2
Multimodal LLMs
330.1LLAVA-OV + cdViews
LLAVA-OV
3D Question Answering via only 2D Vision-Language Models
ICML 2025 · Ablation table
Visual question answering
428MemTree3D
LLaVA-OneVision-7B
Memory Tree Guided Key Frame Querying for Efficient 3D Question Answering
ECCV 2026 · Table 3
Video-language understanding
527.9+ QA samples
LLaVA-3D
MMScan: A Multi-Modal 3D Scene Dataset with Hierarchical Grounded Language Annotations
NeurIPS 2024 · Using VG and QA for tuning tab
Human and embodied modeling
627.8DTC (54%)
LLaVA-OneVision-7B
Zero-shot 3D Question Answering via Voxel-based Dynamic Token Compression
CVPR 2025 · Table 3
Visual question answering
727.3Qwen-3D 7B
Qwen2.5-VL-7B
Qwen-3D: A Generalist 3D Vision-Language Model for Spatial Understanding
ECCV 2026 · Table 1
Multimodal LLMs
827.22Sparse3DPR (Ours, training-free, sparse RGB)Sparse3DPR: Training-Free 3D Hierarchical Scene Parsing and Task-Adaptive Subgraph Reasoning from Sparse RGB Views
AAAI 2026 · Table 2
3D generation
926.98BridgeQA
BLIP
Bridging the Gap between 2D and 3D Visual Question Answering: A Fusion Approach for 3D VQA
AAAI 2024 · Table 1
Visual question answering
1026LLaVA3LLaVA³: Representing 3D Scenes Like a Cubist Painter to Boost 3D Scene Understanding of VLMs
AAAI 2026 · Table 1
3D generation
5 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.