ViewSpatial-Bench

Multi-View Spatial Model 82.1% tops its leaderboard

InternVL2.5-8B 43.2% leads the models we track

SenseNova-SI-1.3-InternVL3-8B 61.3% leads since January 2026

Fading

Official leaderboard Include claimed scores

Models answer 5,712 questions that locate objects from the camera view and from the view of people in the scene. Five task types test egocentric and allocentric reasoning.

Questions were made by automatic generation plus human annotation.

Made by: Dingming Li, Yueting Zhuang et al.
Released: May 2025
Kind: Spatial reasoning and counting
Level: Everyday
Scored: Checked automatically
Answers: Multiple choice
Tasks: All public

Strongest model tested: Gemini 2.0 Flash Feb 2025 #146 on the capability index
Last in a launch post: January 2026, dropped before it saturated
Last model added: October 2026
Scores refreshed: November 12th, 2026

Paper arxiv.org

In the authors’ words

Vision-language models (VLMs) have demonstrated remarkable capabilities in understanding and reasoning about visual content, but significant challenges persist in tasks requiring cross-viewpoint understanding and spatial reasoning. We identify a critical limitation: current VLMs excel primarily at egocentric spatial reasoning (from the camera's perspective) but fail to generalize to allocentric viewpoints when required to adopt another entity's spatial frame of reference. We introduce ViewSpatial-Bench, the first comprehensive benchmark designed specifically for multi-viewpoint spatial localization recognition evaluation across five distinct task types, supported by an automated 3D annotation pipeline that generates precise directional labels. Comprehensive evaluation of diverse VLMs on ViewSpatial-Bench reveals a significant performance disparity: models demonstrate reasonable performance on camera-perspective tasks but exhibit reduced accuracy when reasoning from a human viewpoint. By fine-tuning VLMs on our multi-perspective spatial dataset, we achieve an overall performance improvement of 46.24% across tasks, highlighting the efficacy of our approach. Our work establishes a crucial benchmark for spatial intelligence in embodied AI systems and provides empirical evidence that modeling 3D spatial relationships enhances VLMs' corresponding spatial comprehension capabilities.

Abstract of ViewSpatial-Bench: Evaluating Multi-perspective Spatial Localization in Vision-Language Models (arXiv:2505.21500, CC0)

Entries on its own leaderboard, 7 models we track

Launch post lists it, from 1 lab

System cards report it, from 2 labs

Multi-View Spatial Model tops the leaderboard with 82.1%

Best score per entry from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is InternVL2.5-8B, at 43.2%.

SenseNova-SI-1.3-InternVL3-8B leads with 61.3%, claims included

Best score per model from every source, including labs’ own numbers and claims Epoch AI collected from reports and papers. Labs pick their own settings, so a claim can beat an independent run of the same model.

Bars Table

The newest model with a score here came out in April 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

1 Multi-View Spatial Model zero-shot 82.1%
2 InternVL2.5-8B zero-shot 43.2%
3 InternVL3-14B zero-shot 40.3%
4 Qwen2.5-VL-7B zero-shot 36.9%
5 Qwen2.5-VL-3B zero-shot 35.9%
6 GPT-4o (snapshot unspecified) zero-shot 35%
7 InternVL2.5-2B zero-shot 35%
8 Kimi-VL-Instruct-16B zero-shot 33.6%
9 Gemini 2.0 Flash Feb 2025, zero-shot 32.6%
10 LLaVA-NeXT-Video-7B zero-shot 30.6%
11 Llama-3.2-Vision-11B zero-shot 28.8%
12 LLaVA-OneVision 7B zero-shot 27.5%

Source: Official ViewSpatial-Bench leaderboard

1 SenseNova-SI-1.3-InternVL3-8B 61.3%
2 Nemotron 3.5 Super VL 52%
3 STEP3-VL-10B 46.1%
4 InternVL2.5-8B zero-shot 43.2%
5 InternVL3-14B zero-shot 40.3%
6 Qwen2.5-VL-7B zero-shot 36.9%
7 Qwen2.5-VL-3B zero-shot 35.9%
8 InternVL2.5-2B zero-shot 35%
9 Gemini 2.0 Flash Feb 2025, zero-shot 32.6%
10 LLaVA-OneVision 7B zero-shot 27.5%

Sources: Labs’ model cards and official ViewSpatial-Bench leaderboard

#	Entry	Score	Setting	Date	Source
1	Multi-View Spatial Model	82.1%	zero-shot		Official leaderboard
2	InternVL2.5-8B	43.2%	zero-shot	Dec 2024	Official leaderboard
3	InternVL3-14B	40.3%	zero-shot	Apr 2025	Official leaderboard
4	Qwen2.5-VL-7B	36.9%	zero-shot	Jan 2025	Official leaderboard
5	Qwen2.5-VL-3B	35.9%	zero-shot	Jan 2025	Official leaderboard
6	GPT-4o (snapshot unspecified)	35%	zero-shot		Official leaderboard
7	InternVL2.5-2B	35%	zero-shot	Dec 2024	Official leaderboard
8	Kimi-VL-Instruct-16B	33.6%	zero-shot		Official leaderboard
9	Gemini 2.0 Flash Feb 2025	32.6%	zero-shot	Feb 2025	Official leaderboard
10	LLaVA-NeXT-Video-7B	30.6%	zero-shot		Official leaderboard
11	Llama-3.2-Vision-11B	28.8%	zero-shot		Official leaderboard
12	LLaVA-OneVision 7B	27.5%	zero-shot	Aug 2024	Official leaderboard

#	Model	Score	Setting	Date	Source
1	SenseNova-SI-1.3-InternVL3-8B	61.3%		Jan 2026	Lab model card
2	Nemotron 3.5 Super VL	52%		Oct 2026	Lab model card
3	STEP3-VL-10B	46.1%		Jan 2026	Lab model card
4	InternVL2.5-8B	43.2%	zero-shot	Dec 2024	Official leaderboard
5	InternVL3-14B	40.3%	zero-shot	Apr 2025	Official leaderboard
6	Qwen2.5-VL-7B	36.9%	zero-shot	Jan 2025	Official leaderboard
7	Qwen2.5-VL-3B	35.9%	zero-shot	Jan 2025	Official leaderboard
8	InternVL2.5-2B	35%	zero-shot	Dec 2024	Official leaderboard
9	Gemini 2.0 Flash Feb 2025	32.6%	zero-shot	Feb 2025	Official leaderboard
10	LLaVA-OneVision 7B	27.5%	zero-shot	Aug 2024	Official leaderboard

Labs report scores for 3 models in their cards. Switch the scores above to rank them together.

Number one over time 2 models have held number one with claims included since its release in May 2025

Lab claims count from the day the card came out, independent runs from the model’s release. We draw one row each for the benchmark’s own leaderboard and what labs claim in their cards.

Shanghai AI Lab SenseTime

Own leaderboard Not tested since May 2025

InternVL2.5-8B 43.2%

Lab claims 1 leader

SenseNova-SI-1.3-InternVL3-8B 61.3%

Sources: Official ViewSpatial-Bench leaderboard and labs’ model cards

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

On the benchmark’s own leaderboard Reported by the lab in its model card Best so far

Sources: Official ViewSpatial-Bench leaderboard and labs’ model cards

Labs headline it less often now, last in January 2026

Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.

SenseTime

Date	Launch post	Score	As the post writes it
Jan 8, 2026	SenseNova-SI-1.3-InternVL3-8B	61.3	ViewSpatial

2 system cards report it, from 2 labs

Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.

Date	Document	Score	As the document writes it	Section
Oct 15, 2026	Nemotron 3.5 Super VL technical report	52	ViewSpatial, SpatialClaw harness, with agentic vision training, Table 7	Capabilities
Jan 14, 2026	STEP3-VL-10B technical report	46.1	ViewSpatial, SeRe (Tables 1 and 3). PaCoRe 48.41	Capabilities