ModelsSep 6, 2026 · 21:388 sources

OpenAI's Astra Scores 77.3% on Browser Use Benchmark v2 Versus Opus 5's 50.5%

From huggingnews.com

Image: huggingnews.com
AI;DR

OpenAI's GPT-6 Astra scored 77.3% on Browser Use Benchmark v2 against Opus 5's 50.5%, and 92.9% on WeirdML, matching Fable 5.1's top score. Early tests showed Astra handling browser, desktop, and visual tasks from prompts, rebuilding an apartment display with a telescope feed in about 20 minutes, and solving logic puzzles without code. Some users reported Astra consuming tokens quickly, with one Pro x20 subscriber saying a day of Astra Light use nearly exhausted the plan's weekly usage.

Read the sourceSign in to suggest

Sources

+ 6 more
  1. SUPPORT@npew“Astra just obliterated our hardest benchmark 👀”x.com
  2. SUPPORT@rhyssullivan“20 minutes later it was done and worked great”x.com
  3. SUPPORT@teortaxestex“GPT-6 Astra is the best "vision" model I've seen”x.com
  4. SUPPORT@scaling01“it logically (no code) solved ALL of them”x.com
  5. SUPPORT@htihle“It sets a new individual high score on 6 of the 17 tasks”x.com
  6. SUPPORT@koltregaskes“basically out of my weekly usage on Pro x20 plan”x.com

One ranked edition, every morning

The stories that matter today. Free, cancel anytime.

Telegram @aidr_today