# OpenAI’s GPT-6 Astra drives Toyota Corolla through parking lot course

Digest AI · Robotics & Physical AI · published 2026-09-23T12:01:00Z

Canonical: https://digestai.news/story/openais-gpt-6-astra-drives-toyota-corolla-through-parking-lot-course

## Summary

A new benchmark called **DrivingBench** tested whether frontier language models could control a real Toyota Corolla in a parking lot. Only OpenAI’s **GPT-6 Astra** succeeded, completing a roughly **130-meter course** on its second attempt in **5 minutes 22 seconds**. The other models—**Claude Fable 5.1**, **Grok 4.6**, and **GPT-5.6 Sol**—failed to progress past the halfway mark in any attempt.

The test, created by **Aditya Ramabadran**, **Simon Mahns**, and **Tobias Gessler**, required models to navigate cones using a **comma four** hardware setup with **openpilot**. Models had to interpret camera feeds, adjust steering and speed, and account for latency. **Astra** averaged under **1 mph**, while **Fable 5.1** reached **45%** progress on its third try. Failures often stemmed from misreading lane markers or poor command timing. The researchers noted **Astra** showed in-context learning, adjusting its approach between attempts, while **Grok** and **Sol** struggled with perception and control cadence.

The benchmark is open-source, with all runs published as video traces. The team plans to expand it with more models, harder courses, and varied reasoning efforts. They emphasize the need for safer evaluation methods for real-world AI control.

## Key points

- OpenAI’s **GPT-6 Astra** completed a **130-meter parking lot course** in **5:22**, the only model to finish
- **Claude Fable 5.1** reached **45%** progress on its third attempt, others failed before halfway
- Models used **comma four** + **openpilot** with **MCP tools** for steering, speed, and braking commands

## Why it matters

This is the first real-world test of general-purpose LLMs driving physical vehicles, proving frontier models can handle messy, dynamic tasks beyond simulation. It highlights strengths in **Astra’s** physical reasoning but also exposes gaps in perception and safety.

## Sources

1. [Driving Bench: GPT-6 Astra Becomes First Model To Drive A Real Car Through A Course](https://officechai.com/ai/driving-bench-gpt-6-astra-becomes-first-model-to-drive-a-real-car-through-a-course) (officechai.com, 2026-09-23)

## Cite

Digest AI, "OpenAI’s GPT-6 Astra drives Toyota Corolla through parking lot course", 23 September 2026, https://digestai.news/story/openais-gpt-6-astra-drives-toyota-corolla-through-parking-lot-course

---

Written by Digest AI's editorial model from the linked sources; the sources are the record. Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse
JSON: https://digestai.news/story/openais-gpt-6-astra-drives-toyota-corolla-through-parking-lot-course.json
