GPT-6 Astra drives a car in leaderboard tests
A leaderboard released on Hacker News shows that the multimodal model GPT-6 Astra completed a simulated driving course with 100 % progress in 5 minutes 22 seconds. The test allowed up to three attempts in a single continuous chat, recording distance, speed, finish time and token usage for each run. In the same benchmark, Anthropic’s Claude Fable 5.1 reached 45 % of the course, while Grok 4.6…
Key points
- GPT-6 Astra completed the driving course with 100% progress in 5:22, outperforming other models.
- Claude Fable 5.1 reached 45%, Grok 4.6 reached 11%, GPT-5.6 reached 6% on the same test.
- Leaderboard records distance, speed, finish time, token usage and cost for up to three attempts per model.
The table also lists token counts and list‑price costs for each attempt, though exact numbers are not shown. GPT‑6 Astra’s performance suggests it can maintain lane position within a 4‑meter corridor while navigating the track, a step toward autonomous‑driving capabilities for large language models. The results are presented without independent verification, and the source does not name the organization behind GPT‑6 Astra.
Model pages: GPT-6 Astra → · Claude Fable 5.1 →
GPT-6 Astra has gained the ability to drive a car
drivingbench.com · 20 September 2026
Loading the full article…
This text was published by drivingbench.com. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
Coverage and discussion
1source- Hacker News discussion · 213 pointsnews.ycombinator.com
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Research
All →- Anthropic's Claude begins 'dreaming' to reorganize memory in tests · 1 src
- John Langford quits Microsoft, declares AI Frontiers over · 1 src
- Anthropic releases study on Claude usage in scientific research · 3 src
- MIT and BASF release CheMeleon for low-data molecular prediction · 2 src
- Anthropic tests Claude agents in EVE Online trading and sabotage detection · 1 src
Comments
via GitHub Discussions