TriPhysBench
Models answer quantitative questions about physical events in videos. Some questions require measuring an object or looking up its specifications.
The authors compare six models on 410 questions.
- Made by
- Renmin University of China and collaborators
- Released
- November 2026
- Kind
- Embodied reasoning
- Level
- School
- Scored
- Checked automatically
- Answers
- Short answer
- Measures
- Capability
- Subject
- Physical sciences
- Language
- English
- Interaction
- Agentic
- Environment
- APIs and tools, Code sandbox
- Refresh
- One fixed set
- Run by
- A host
- Strongest model tested
- GPT-5.5 #21 on the capability index
- Last model added
- July 2026
- Scores refreshed
- November 20th, 2026
Gemini 3.6 Flash leads with 50.4% among models released by July 2026 Gemini 3.6 Flash leads with 50.4% among models released by July 2026
Best score per entry from the benchmark's paper, read on November 20th, 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. This view shows only models we track, each at its best.
The newest model with a score here came out in July 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
- 1
Gemini 3.6 Flash 50.4%
- 2 GPT-5.5 44.4%
- 3
Qwen3.6 27B 43.8%
- 4
Qwen3.5-9B 42%
- 5
Qwen3.5-4B 39.2%
- 6 InteractiveOmni-4B 32.6%
Source: The TriPhysBench paper
- 1
Gemini 3.6 Flash 50.4%
- 2 GPT-5.5 44.4%
- 3
Qwen3.6 27B 43.8%
- 4
Qwen3.5-9B 42%
- 5
Qwen3.5-4B 39.2%
Source: The TriPhysBench paper
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 50.4% | Jul 2026 | Official leaderboard | |
| 2 | GPT-5.5 | 44.4% | Apr 2026 | Official leaderboard | |
| 3 | | 43.8% | Apr 2026 | Official leaderboard | |
| 4 | | 42% | Feb 2026 | Official leaderboard | |
| 5 | | 39.2% | Feb 2026 | Official leaderboard | |
| 6 | InteractiveOmni-4B | 32.6% | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 50.4% | Jul 2026 | Official leaderboard | |
| 2 | GPT-5.5 | 44.4% | Apr 2026 | Official leaderboard | |
| 3 | | 43.8% | Apr 2026 | Official leaderboard | |
| 4 | | 42% | Feb 2026 | Official leaderboard | |
| 5 | | 39.2% | Feb 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 TriPhysBench 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 TriPhysBench 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 Robotics
- ERQA In use 400 multiple-choice visual questions from robot and egocentric footage on spatial, trajectory, action and state reasoning. Accuracy.
Gemini Robotics ER 2 78.5%
- SimplerEnv In use Real-to-sim replicas of Google Robot and WidowX/Bridge setups. Policies scored by success rate on pick, move, drawer and Bridge tasks. Praxis-1 80%
- RoboLab In use Zero-shot DROID Franka policies in Isaac Lab on 120 tabletop tasks at three instruction levels, 10 episodes each. Success rate. Praxis-1 46.2%
- RoboDojo In use Generalist manipulation policies run 42 Isaac Sim tasks and 18 remote real-robot tasks. Scored by task score and success rate. GPT-6 Astra 27.6%
- Progress classification (Google DeepMind) In use Models assign each frame of a robot video to one of five task-progress bands from 0 to 100 percent. Scored by accuracy.
Gemini Robotics ER 2 57.4%
- Human proximity (Google DeepMind robotics) In use Models judge from robot camera views whether a human is within 1 meter, a spatial awareness check for safe robot behavior. Scored by accuracy.
Gemini Robotics ER 2 93%
Data audit
9 of 10 checks pass, 1 warning. Look first at the Task access check. We don’t know whether the tasks are public. Every benchmark
Catalog record 1 warning 1 pass
- Warning Task access We don’t know whether the tasks are public.
- Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
- Pass Independent runs Scores from the benchmark’s paper (6 entries).
- Pass Models scored 5 models we track, plus 1 other entry on its leaderboard.
- Pass Newest score The newest is from July 2026, for Gemini 3.6 Flash.
- Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.5 (#21).
- Pass Headroom The best is 50.4%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on arxiv.org: 5 rows for models we track and 1 other entry at the last read.
- Pass Last read November 20th, 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.