MESA
Robot policies follow instructions as object locations, categories, and task sequences change. The authors compare seven methods on single-arm and two-arm tasks.
The authors call the bimanual suite BiMESA.
- Made by
- Georgia Institute of Technology
- Released
- November 2026
- Kind
- Simulated robots
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Agent task
- Tasks
- All public
- Measures
- Capability
- Language
- English
- In and out
- Text, Images, Time series → Actions
- Interaction
- Agentic
- Environment
- A simulator
- Refresh
- One fixed set
- Run by
- Anyone
- Last model added
- December 2025
- Scores refreshed
- November 20th, 2026
π0.5 leads with 50.4% among models released by December 2025 π0.5 leads with 50.4% among models released by December 2025
Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. The headline averages the published single-arm MESA and bimanual BiMESA success rates. 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 December 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 π0.5 50.4%
- 2 π0-FAST 37.4%
- 3 π0 30.6%
- 4
GR00T N1.6 27.8%
- 5 PG-FM 26.6%
- 6 DP 5.85%
Source: Official MESA leaderboard
- 1 π0.5 50.4%
- 2 π0-FAST 37.4%
- 3 π0 30.6%
- 4
GR00T N1.6 27.8%
Source: Official MESA leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | π0.5 | 50.4% | Apr 2025 | Official leaderboard | |
| 2 | π0-FAST | 37.4% | Jan 2025 | Official leaderboard | |
| 3 | π0 | 30.6% | Oct 2024 | Official leaderboard | |
| 4 | | 27.8% | Dec 2025 | Official leaderboard | |
| 5 | PG-FM | 26.6% | Official leaderboard | ||
| 6 | DP | 5.85% | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | π0.5 | 50.4% | Apr 2025 | Official leaderboard | |
| 2 | π0-FAST | 37.4% | Jan 2025 | Official leaderboard | |
| 3 | π0 | 30.6% | Oct 2024 | Official leaderboard | |
| 4 | | 27.8% | Dec 2025 | Official leaderboard |
Rankings by part
We read 12 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as MESA-ID, MESA-Spatial and MESA-Instance. π0.5 leads 11 of them, and GR00T N1.6 leads the other one. Arrows show how far a model moved from its overall rank among the models we track.
MESA-ID
4 models we track have a score here.
- 1 π0.5 62.8%
- 2 π0-FAST 46.5%
- 3 π0 43.1%
- 4
GR00T N1.6 25%
MESA-Spatial
4 models we track have a score here.
- 1 π0.5 68.6%
- 2 π0-FAST 53.6%
- 3 π0 42.2%
- 4
GR00T N1.6 23.2%
MESA-Instance
4 models we track have a score here.
- 1 π0.5 55.4%
- 2 π0-FAST 39.1%
- 3 π0 34%
- 4
GR00T N1.6 22.7%
MESA-Composite
4 models we track have a score here.
- 1 π0.5 46.7%
- 2 π0-FAST 28.9%
- 3 π0 27.5%
- 4
GR00T N1.6 15%
MESA-Category
4 models we track have a score here.
- 1 π0.5 26.4%
- 2
GR00T N1.6 ↑2 14.1%
- 3 π0 11.6%
- 4 π0-FAST ↓2 8.40%
MESA Avg.
4 models we track have a score here.
- 1 π0.5 52%
- 2 π0-FAST 35.3%
- 3 π0 31.7%
- 4
GR00T N1.6 20%
BiMESA-ID
4 models we track have a score here.
- 1 π0.5 68.7%
- 2 π0-FAST 53.7%
- 3
GR00T N1.6 ↑1 47%
- 4 π0 ↓1 45.9%
BiMESA-Spatial
4 models we track have a score here.
- 1 π0.5 70.5%
- 2 π0-FAST 58.1%
- 3 π0 49.2%
- 4
GR00T N1.6 46.2%
BiMESA-Instance
4 models we track have a score here.
- 1 π0.5 48.1%
- 2 π0-FAST 36.7%
- 3
GR00T N1.6 ↑1 36.7%
- 4 π0 ↓1 23.9%
BiMESA-Composite
4 models we track have a score here.
- 1 π0.5 37.5%
- 2 π0-FAST 28.8%
- 3
GR00T N1.6 ↑1 22.1%
- 4 π0 ↓1 18.4%
BiMESA-Category
4 models we track have a score here.
- 1
GR00T N1.6 ↑3 26.1%
- 2 π0-FAST 19.8%
- 3 π0.5 ↓2 18.6%
- 4 π0 ↓1 10.6%
BiMESA Avg.
4 models we track have a score here.
- 1 π0.5 48.7%
- 2 π0-FAST 39.4%
- 3
GR00T N1.6 ↑1 35.6%
- 4 π0 ↓1 29.6%
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official MESA leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official MESA 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 Newest score check. The newest came 11 months ago, in December 2025, for GR00T N1.6. 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 official leaderboard (6 entries).
- Pass Models scored 4 models we track, plus 2 other entries on its leaderboard.
- Warning Newest score The newest came 11 months ago, in December 2025, for GR00T N1.6.
- Note Frontier models None of its models are on the capabilities index.
- 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 pairlab.github.io: 4 rows for models we track and 2 other entries 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.