SpaceDG-Bench
Models answer spatial questions under clean images and nine forms of visual degradation. Choice and binary answers use accuracy, while numerical answers use relative-error thresholds. The test has 1,102 verified questions from scenes excluded from the training set.
This record covers SpaceDG-Bench, not the million-question SpaceDG training collection. The benchmark renders clean and degraded versions of the same questions, so its roughly ten thousand instances are not independent questions.
- Released
- May 2026
- Kind
- Spatial reasoning and counting
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Short answer
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Single turn
- Items from
- Generated
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Vision
- Strongest model tested
- GPT-5.4 #28 on the capability index
- Last model added
- April 2026
- Scores refreshed
- November 18th, 2026
In the authors’ words
Multimodal Large Language Models (MLLMs) have made rapid progress in spatial intelligence, yet existing spatial reasoning benchmarks largely assume pristine visual inputs and overlook the degradations that commonly occur in real-world deployment, such as motion blur, low light, adverse weather, lens distortion, and compression artifacts. This raises a fundamental question: how robust is the spatial intelligence of current MLLMs when visual observations are imperfect? To answer this question, we introduce SpaceDG, the first large-scale dataset for degradation-aware spatial understanding. It is constructed with a physically grounded degradation synthesis engine that embeds degradation formation process into 3D Gaussian Splatting (3DGS) rendering, enabling realistic simulation of nine degradation types. The resulting dataset contains approximately 1M QA pairs from nearly 1,000 indoor scenes. We further introduce SpaceDG-Bench, an human-verified benchmark with 1,102 questions spanning 11 reasoning categories and 9 visual degradation types, yielding over 10K VQA instances. Evaluating 25 open- and closed-source MLLMs reveals that visual degradations consistently and substantially impair spatial reasoning, exposing a critical robustness gap. Finally, we show that finetuning on SpaceDG markedly improves degradation robustness and can even surpass human performance under degraded conditions without any performance drop on clean images, highlighting the promise of degradation-aware training for robust spatial intelligence.
SpaceDG-SFT-Qwen3-VL-8B-Instruct tops the leaderboard with 66.1% Gemini 3.1 Pro leads with 56.7% among models released by April 2026
Best score per entry from the benchmark's paper, read on November 18th, 2026. A paper’s table doesn’t add newer models. 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 Gemini 3.1 Pro, at 56.7%. This view shows only models we track, each at its best.
The newest model with a score here came out in April 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 66.1 SpaceDG-SFT-Qwen3-VL-8B-Instruct
- 65.2 SpaceDG-SFT-InternVL-3.5-8B
- 56.7
Gemini 3.1 Pro
- 53.3 SenseNova-SI-InternVL3-8B
- 48.8
Gemini 3.1 Flash-Lite
- 47.7
InternVL3.5-38B
- 46.2 GPT-5.4
- 46.0
Qwen 3.6 Plus
- 45.2 RynnBrain-8B
- 45.2
Qwen3-VL-32B-Instruct
- 45.1 ACE-Brain-0-8B
- 44.2 Claude Sonnet 4.6
- 43.8
Qwen3.5-27B
- 42.4
Qwen 3.6 35B-A3B
- 42.1
Qwen3-VL-8B-Instruct
- 41.7
Qwen3.5-9B
- 41.4
InternVL3.5-8B
- 41.1 VST-7B
- 40.4
Qwen3.5-4B
- 40.1
InternVL3-8B
- 39.7
Qwen3-VL-4B-Instruct
- 35.6 Grok 4.1 Fast
- 34.3
Llama 4 Maverick
- 34.0 Kimi-VL-A3B-Instruct
- 33.8
LLaVA-OneVision 7B (SI)
- 31.0
Gemma 4 26B A4B
- 26.9 Cambrian-S-7B
- 1 SpaceDG-SFT-Qwen3-VL-8B-Instruct 9918 degraded instances 66.1%
- 2 SpaceDG-SFT-InternVL-3.5-8B 9918 degraded instances 65.2%
- 3
Gemini 3.1 Pro 9918 degraded instances 56.7%
- 4 SenseNova-SI-InternVL3-8B 9918 degraded instances 53.3%
- 5
Gemini 3.1 Flash-Lite 9918 degraded instances 48.8%
- 6
InternVL3.5-38B 9918 degraded instances 47.7%
- 7 GPT-5.4 9918 degraded instances 46.2%
- 8
Qwen 3.6 Plus 9918 degraded instances 46%
- 9 RynnBrain-8B 9918 degraded instances 45.2%
- 10
Qwen3-VL-32B-Instruct 9918 degraded instances 45.2%
- 11 ACE-Brain-0-8B 9918 degraded instances 45.1%
- 12 Claude Sonnet 4.6 9918 degraded instances 44.2%
- 13
Qwen3.5-27B 9918 degraded instances 43.8%
- 14
Qwen 3.6 35B-A3B 9918 degraded instances 42.4%
- 15
Qwen3-VL-8B-Instruct 9918 degraded instances 42.1%
- 16
Qwen3.5-9B 9918 degraded instances 41.7%
- 17
InternVL3.5-8B 9918 degraded instances 41.4%
- 18 VST-7B 9918 degraded instances 41.1%
- 19
Qwen3.5-4B 9918 degraded instances 40.4%
- 20
InternVL3-8B 9918 degraded instances 40.1%
- 21
Qwen3-VL-4B-Instruct 9918 degraded instances 39.7%
- 22 Grok 4.1 Fast 9918 degraded instances 35.6%
- 23
Llama 4 Maverick 9918 degraded instances 34.3%
- 24 Kimi-VL-A3B-Instruct 9918 degraded instances 34%
- 25
LLaVA-OneVision 7B (SI) 9918 degraded instances 33.8%
- 26
Gemma 4 26B A4B 9918 degraded instances 31%
- 27 Cambrian-S-7B 9918 degraded instances 26.9%
Source: The SpaceDG-Bench paper
- 56.7
Gemini 3.1 Pro
- 48.8
Gemini 3.1 Flash-Lite
- 47.7
InternVL3.5-38B
- 46.2 GPT-5.4
- 46.0
Qwen 3.6 Plus
- 45.2
Qwen3-VL-32B-Instruct
- 44.2 Claude Sonnet 4.6
- 43.8
Qwen3.5-27B
- 42.4
Qwen 3.6 35B-A3B
- 42.1
Qwen3-VL-8B-Instruct
- 41.7
Qwen3.5-9B
- 41.4
InternVL3.5-8B
- 40.4
Qwen3.5-4B
- 40.1
InternVL3-8B
- 39.7
Qwen3-VL-4B-Instruct
- 35.6 Grok 4.1 Fast
- 34.3
Llama 4 Maverick
- 34.0 Kimi-VL-A3B-Instruct
- 33.8
LLaVA-OneVision 7B (SI)
- 31.0
Gemma 4 26B A4B
- 1
Gemini 3.1 Pro 9918 degraded instances 56.7%
- 2
Gemini 3.1 Flash-Lite 9918 degraded instances 48.8%
- 3
InternVL3.5-38B 9918 degraded instances 47.7%
- 4 GPT-5.4 9918 degraded instances 46.2%
- 5
Qwen 3.6 Plus 9918 degraded instances 46%
- 6
Qwen3-VL-32B-Instruct 9918 degraded instances 45.2%
- 7 Claude Sonnet 4.6 9918 degraded instances 44.2%
- 8
Qwen3.5-27B 9918 degraded instances 43.8%
- 9
Qwen 3.6 35B-A3B 9918 degraded instances 42.4%
- 10
Qwen3-VL-8B-Instruct 9918 degraded instances 42.1%
- 11
Qwen3.5-9B 9918 degraded instances 41.7%
- 12
InternVL3.5-8B 9918 degraded instances 41.4%
- 13
Qwen3.5-4B 9918 degraded instances 40.4%
- 14
InternVL3-8B 9918 degraded instances 40.1%
- 15
Qwen3-VL-4B-Instruct 9918 degraded instances 39.7%
- 16 Grok 4.1 Fast 9918 degraded instances 35.6%
- 17
Llama 4 Maverick 9918 degraded instances 34.3%
- 18 Kimi-VL-A3B-Instruct 9918 degraded instances 34%
- 19
LLaVA-OneVision 7B (SI) 9918 degraded instances 33.8%
- 20
Gemma 4 26B A4B 9918 degraded instances 31%
Source: The SpaceDG-Bench paper
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | SpaceDG-SFT-Qwen3-VL-8B-Instruct | 66.1% | 9918 degraded instances | Official leaderboard | |
| 2 | SpaceDG-SFT-InternVL-3.5-8B | 65.2% | 9918 degraded instances | Official leaderboard | |
| 3 | | 56.7% | 9918 degraded instances | Feb 2026 | Official leaderboard |
| 4 | SenseNova-SI-InternVL3-8B | 53.3% | 9918 degraded instances | Official leaderboard | |
| 5 | | 48.8% | 9918 degraded instances | Mar 2026 | Official leaderboard |
| 6 | | 47.7% | 9918 degraded instances | Aug 2025 | Official leaderboard |
| 7 | GPT-5.4 | 46.2% | 9918 degraded instances | Mar 2026 | Official leaderboard |
| 8 | | 46% | 9918 degraded instances | Mar 2026 | Official leaderboard |
| 9 | RynnBrain-8B | 45.2% | 9918 degraded instances | Official leaderboard | |
| 10 | | 45.2% | 9918 degraded instances | Oct 2025 | Official leaderboard |
| 11 | ACE-Brain-0-8B | 45.1% | 9918 degraded instances | Official leaderboard | |
| 12 | Claude Sonnet 4.6 | 44.2% | 9918 degraded instances | Feb 2026 | Official leaderboard |
| 13 | | 43.8% | 9918 degraded instances | Feb 2026 | Official leaderboard |
| 14 | | 42.4% | 9918 degraded instances | Apr 2026 | Official leaderboard |
| 15 | | 42.1% | 9918 degraded instances | Oct 2025 | Official leaderboard |
| 16 | | 41.7% | 9918 degraded instances | Feb 2026 | Official leaderboard |
| 17 | | 41.4% | 9918 degraded instances | Aug 2025 | Official leaderboard |
| 18 | VST-7B | 41.1% | 9918 degraded instances | Official leaderboard | |
| 19 | | 40.4% | 9918 degraded instances | Feb 2026 | Official leaderboard |
| 20 | | 40.1% | 9918 degraded instances | Apr 2025 | Official leaderboard |
| 21 | | 39.7% | 9918 degraded instances | Oct 2025 | Official leaderboard |
| 22 | Grok 4.1 Fast | 35.6% | 9918 degraded instances | Nov 2025 | Official leaderboard |
| 23 | | 34.3% | 9918 degraded instances | Apr 2025 | Official leaderboard |
| 24 | Kimi-VL-A3B-Instruct | 34% | 9918 degraded instances | Apr 2025 | Official leaderboard |
| 25 | | 33.8% | 9918 degraded instances | Aug 2024 | Official leaderboard |
| 26 | | 31% | 9918 degraded instances | Apr 2026 | Official leaderboard |
| 27 | Cambrian-S-7B | 26.9% | 9918 degraded instances | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 56.7% | 9918 degraded instances | Feb 2026 | Official leaderboard |
| 2 | | 48.8% | 9918 degraded instances | Mar 2026 | Official leaderboard |
| 3 | | 47.7% | 9918 degraded instances | Aug 2025 | Official leaderboard |
| 4 | GPT-5.4 | 46.2% | 9918 degraded instances | Mar 2026 | Official leaderboard |
| 5 | | 46% | 9918 degraded instances | Mar 2026 | Official leaderboard |
| 6 | | 45.2% | 9918 degraded instances | Oct 2025 | Official leaderboard |
| 7 | Claude Sonnet 4.6 | 44.2% | 9918 degraded instances | Feb 2026 | Official leaderboard |
| 8 | | 43.8% | 9918 degraded instances | Feb 2026 | Official leaderboard |
| 9 | | 42.4% | 9918 degraded instances | Apr 2026 | Official leaderboard |
| 10 | | 42.1% | 9918 degraded instances | Oct 2025 | Official leaderboard |
| 11 | | 41.7% | 9918 degraded instances | Feb 2026 | Official leaderboard |
| 12 | | 41.4% | 9918 degraded instances | Aug 2025 | Official leaderboard |
| 13 | | 40.4% | 9918 degraded instances | Feb 2026 | Official leaderboard |
| 14 | | 40.1% | 9918 degraded instances | Apr 2025 | Official leaderboard |
| 15 | | 39.7% | 9918 degraded instances | Oct 2025 | Official leaderboard |
| 16 | Grok 4.1 Fast | 35.6% | 9918 degraded instances | Nov 2025 | Official leaderboard |
| 17 | | 34.3% | 9918 degraded instances | Apr 2025 | Official leaderboard |
| 18 | Kimi-VL-A3B-Instruct | 34% | 9918 degraded instances | Apr 2025 | Official leaderboard |
| 19 | | 33.8% | 9918 degraded instances | Aug 2024 | Official leaderboard |
| 20 | | 31% | 9918 degraded instances | Apr 2026 | Official leaderboard |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The SpaceDG-Bench paper
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official SpaceDG-Bench leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Vision
- MMMU In use 11.5K college-level multimodal questions across 30 subjects. Multiple-choice and open accuracy.
Qwen 3.6 Plus 86%
- MMMU-Pro In use Harder MMMU: college-level multimodal questions with more options and vision-only inputs to remove text shortcuts. GPT-5.6 Sol 84.6%
- CharXiv In use Chart understanding on real arXiv figures. Reasoning questions require synthesizing information from complex charts. Claude Fable 5 93.5%
- MathVista In use Math reasoning in visual contexts (charts, geometry, puzzles) from 31 datasets. Answer accuracy.
Seed2.1 Pro 90.7%
- MATH-Vision (MathVision) In use 3,040 competition math problems with visual context. Answer accuracy.
Qwen 3.8 Max 97.7%
- MMStar In use Models answer 1,500 human-curated image questions that cannot be solved without the image, across 6 core capabilities. Scored by accuracy.
Qwen 3.8 Max 85.9%
Data audit
9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 7 months ago, in April 2026, for Qwen 3.6 35B-A3B. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
- Pass Independent runs Scores from the benchmark’s paper (27 entries).
- Pass Models scored 20 models we track, plus 7 other entries on its leaderboard.
- Warning Newest score The newest came 7 months ago, in April 2026, for Qwen 3.6 35B-A3B.
- Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.4 (#28).
- Pass Headroom The best is 56.7%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on arxiv.org: 22 rows for models we track and 7 other entries at the last read.
- Pass Last read November 18th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.