GPT-6.1 Sol
Frontier LLM evaluated as a robot controller on AA-Robotics, where it runs as a coding agent in Codex at high effort.
OpenAI · Sep 29, 2026 · LLM controllerReport an error
41.7%Dojo-RC · #7 of 13
64.6AA-Robotics · #5 of 9
2Benchmarks reported
Sep 29, 2026Released
Benchmark results2
RoboDojo-RC Tier 1Real robot
- Official leaderboard ↗Oct 2, 2026· high effort, fast tier· n = 120
#7 of 1341.7%
AA-Robotics v0.1Real robot
- Official leaderboard ↗Oct 14, 2026· v0.1· n = 8
Score by task
Press the red button100Place the crocodile toy in the rectangle100Pour the drink into the cup0Tidy the blocks into the box100Complete the shape matching challenge100Write hello world100Build the tallest tower0Measure the box17
#5 of 964.6
Not reported on 15 benchmarks
Speed per task
| Setting | Per task | Per planner call | Calls | Cost | Source |
|---|---|---|---|---|---|
| RoboDojo-RC Tier 1, bimanual YAM, 6 tasks × 20 trials · GPT-6.1 Sol API (high effort, fast tier) | 4 min | – | – | USD 2.15 | Oct 2, 2026 ↗ |
Wrong or missing number? Report an error