Today for AI
HOT RADAR
agentHEAT 12.7°

AI CURATED HOT STORY · 10/22/2026

Study Finds GPT-6-Astra Saturates Robot Benchmarks but Fails Physical Verification

1 reports archived1 independent sourcesupdated 10/22/2026, 13:56:00
Synthesis & Latest Updates
1 Sources Cross-Validated

A new paper evaluates GPT-6-Astra's performance in embodied intelligence tasks, finding it nearly saturates simulation benchmarks like LIBERO but fails in real-world robot deployment due to an inability to verify physical success and prohibitive costs. The study highlights the limitations of using general-purpose models as robot controllers: while capable of multi-task handling, they struggle to recover from irreversible errors and consume approximately 9 million tokens per long-horizon task, blocking practical adoption.

LATEST/A new paper evaluates GPT-6-Astra's performance in embodied intelligence tasks, finding it nearly saturates simulation benchmarks like LIBERO but fails in real-world robot deployment due to an inability to verify physical success and prohibitive costs. The study highlights the limitations of using general-purpose models as robot controllers: while capable of multi-task handling, they struggle to recover from irreversible errors and consume approximately 9 million tokens per long-horizon task, blocking practical adoption.

HEAT TRENDHourly heat curve

3 fully observed hours
Now
12.7
Peak
12.914:00
24h change
–

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

TIMELINECoverage timeline

Total 1 reports · Latest first
  1. Rohan Paul (@rohanpaul_ai)T2·78 pts
    • GPT-6-Astra performs exceptionally well on simulation benchmarks like RoboDojo (81.9%) and DexJoCo (87.3%), nearing saturation.
    • In real-world robot scenarios, the model fails to effectively verify physical execution results and struggles to recover from irreversible errors.
    • A single LIBERO-Ultra-Long task consumes ~9 million tokens and nearly 2 hours of compute time, exceeding practical cost thresholds.