VagueBench
Agents follow paired instructions in indoor environments. The authors compare performance when a clear instruction and a vague instruction describe the same task.
- Made by
- Yang Qu and colleagues
- Released
- April 2026
- Kind
- Embodied reasoning
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Agent task
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Agentic
- Environment
- AI2-THOR
- Refresh
- One fixed set
- Run by
- Anyone
- Last model added
- April 2025
- Scores refreshed
- November 20th, 2026
GPT-4o (snapshot unspecified) tops the leaderboard with 49.1% Qwen2.5-VL-72B leads with 43.5% among models released by April 2025
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. Every run is in the next chart. 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 Qwen2.5-VL-72B, at 43.5%. This view shows only models we track, each at its best.
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:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 1 GPT-4o (snapshot unspecified) Original instructions 49.1%
- 2
Qwen2.5-VL-72B Original instructions 43.5%
- 3
InternVL3-78B Original instructions 38.5%
- 4 Qwen2-VL-72B-Instruct Original instructions 35.4%
Source: The VagueBench paper
Source: The VagueBench paper
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-4o (snapshot unspecified) | 49.1% | Original instructions | Official leaderboard | |
| 2 | | 43.5% | Original instructions | Jan 2025 | Official leaderboard |
| 3 | | 38.5% | Original instructions | Apr 2025 | Official leaderboard |
| 4 | Qwen2-VL-72B-Instruct | 35.4% | Original instructions | Aug 2024 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 43.5% | Original instructions | Jan 2025 | Official leaderboard |
| 2 | | 38.5% | Original instructions | Apr 2025 | Official leaderboard |
| 3 | Qwen2-VL-72B-Instruct | 35.4% | Original instructions | Aug 2024 | Official leaderboard |
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: The VagueBench paper
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| Original instructions | 43.5% | ||||
| Original instructions | 38.5% | ||||
| Qwen2-VL-72B-Instruct | Original instructions | 35.4% | |||
| Vague instructions | 31.6% | ||||
| Qwen2-VL-72B-Instruct | Vague instructions | 29.9% | |||
| Vague instructions | 28.4% |
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: The VagueBench paper
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| Original instructions | 43.5% | ||||
| Original instructions | 38.5% | ||||
| Qwen2-VL-72B-Instruct | Original instructions | 35.4% | |||
| Vague instructions | 31.6% | ||||
| Qwen2-VL-72B-Instruct | Vague instructions | 29.9% | |||
| Vague instructions | 28.4% |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The VagueBench 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 VagueBench 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
8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 19 months ago, in April 2025, for InternVL3-78B. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
- Pass Independent runs Scores from the benchmark’s paper (4 entries).
- Note Models scored Only 3 models we track, plus 1 other entry on its leaderboard.
- Warning Newest score The newest came 19 months ago, in April 2025, for InternVL3-78B.
- Note Frontier models None of its models are on the capabilities index.
- Pass Headroom The best is 43.5%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on arxiv.org: 6 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.