AI Paper Atlas

Research map / Benchmarks / MMSI-Bench

MMSI-Bench leaderboard

11 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on MMSI-Bench. Best accuracy: 43.3 (SenseNova-SI-InternVL3-8B, CVPR 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.

MMSI-Bench · accuracy

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

#accuracy ↑MethodPaperTopic
143.3SenseNova-SI-InternVL3-8B
InternVL3-8B
Scaling Spatial Intelligence with Multimodal Foundation Models
CVPR 2026 · Table 1
Multimodal LLMs
238.9MV-STRIDE-SFT(Full)
Qwen3-VL-8B-Instruct
MV-STRIDE: Enabling MLLMs to Master Multi-View Spatial Reasoning via Hierarchical Capability Modeling
ECCV 2026 · Table 1
Multimodal LLMs
335.4ViewFusion (SFT+RL)
Qwen3-VL-4B
ViewFusion: Structured Spatial Thinking Chains for Multi-View Reasoning
ECCV 2026 · Table 1
Multimodal LLMs
434.8VST-7B-RL
Qwen2.5-VL-7B
Visual Spatial Tuning
ECCV 2026 · Main results table
Multimodal LLMs
533.34D-RGPT-8B
NVILA-Lite-8B
4D-RGPT: Toward Region-level 4D Understanding via Perceptual Distillation
CVPR 2026 · Table 2
Video-language understanding
632.6Qwen3-VL-8B + Holi-Spatial-4M
Qwen3-VL-8B
Holi-Spatial: Evolving Video Streams into Holistic 3D Spatial Intelligence
ICML 2026 · Table 3
3D generation
731.7PolyV
Qwen2.5-VL-7B
Modeling Cross-vision Synergy for Unified Large Vision Model
CVPR 2026 · Table 1
Multimodal LLMs
831.4STAR-R1
Qwen2.5-VL-7B
STAR-R1: Multi-View Spatial TrAnsformation Reasoning by Reinforcing Multimodal LLMs
CVPR 2026 · Table 2
Multimodal LLMs
930.7FactoSR-8B-RL
Qwen3-VL-8B-Instruct
Unfold The World: Factorize 4D Properties in Reinforcing Spatial Understanding
ECCV 2026 · Table 2
Multimodal LLMs
1030.2ViLaSR
Qwen2.5-VL-7B
Reinforcing Spatial Reasoning in Vision-Language Models with Interwoven Thinking and Visual Drawing
NeurIPS 2025 · Table: Performance comparison
Multimodal LLMs
1 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.