OpenAI says GPT-6 Astra underperformed on framed safety tasks
Under a safety framing Astra finished about 10 percent of criteria versus 16 percent on general research tasks.
1
Stories per day · 30 days
7 RSS
Relatedopenai9benchmark3codex3customer-story2robotics2agentic-systems1ai-agents1ai-cloning1alignment1amp1anthropic1benchmarks1benefit-street-partners1browser-agent1cad1chatgpt1chrome1claude-opus1
77 stories, newest first · one card per story
Stories tagged #gpt-6-astra
Chinese labs reportedly distill OpenAI GPT-6 Astra outputs
The Information says DeepSeek and Moonshot used Astra, including hidden reasoning, to train Kimi K3.1 Pro and V4.1-Flash 1017.
Gym-Anything-RT benchmarks vision-language agents in real time
GPT-6 Astra passed 9.0% of 100 real-time minigames, rising to 75.0% when the environment was paused.
Greptile uses GPT-6 Astra to test apps beyond code review
The startup also wires its knowledge base to Codex over MCP and cites Daybreak for customer defenses.
GPT-6-Astra evaluated on EBench for direct and policy-harness robot control
Direct control leads on short tasks but lags on long-horizon work; calling a VLA or WAM matches top policies at lower token cost.
Study finds GPT-6 Astra useful but unreliable for engineering CAD
Tests cover model understanding, generation, editing, and evaluation, with recurring failures in freeform geometry, parameters, and mating.
ChatGPT becomes a full Chrome browser assistant that reads and acts on your open tabs
OpenAI's ChatGPT Chrome extension can now read open tabs and act on them with Computer Use.
Benefit Street Partners answers questions faster with Codex
Benefit Street Partners uses OpenAI Codex and GPT-6 Astra to speed credit underwriting analysis.
Video Game Reverse Engineering Community Shuts Down Because AI Has Made It Too Easy
decomp.dev is shutting down after AI made game decompilation and cloning trivial, overwhelming human-led projects.
7
GPT-6-Astra solves 49 of 50 physics research problems at $2.40 apiece
OpenAI's GPT-6-Astra solved 49 of 50 routine theoretical-physics problems at about $2.40 each.
12
Authors say GPT-6-Astra has competence, not practical intelligence
Authors report strong benchmark results with in-context learning, but real robots still face costly demos and time limits.
Why do LLM pelicans look like this?
Design Creativity Bench finds frontier models lag human designers, with GPT-6 Astra scoring 65.5 versus 92.3.
Amp puts AI agents to work beyond code with OpenAI models
Coding-agent startup Amp says OpenAI models, including GPT-6 Astra, power most of its usage.
Higgsfield takes AI video from 1 minute to over 10
Higgsfield says GPT-6 Astra and GPT-Image-2.5 now support AI videos longer than 10 minutes.
4
Asana cuts model costs 76x in browser tests with GPT-6.1 Sol
Asana cut browser-agent model cost 76x and runtime 5x by optimizing GPT-6.1 Sol with GPT-6 Astra.
Pollo AI turns creative ideas into campaigns with OpenAI
Pollo AI’s agent uses OpenAI’s GPT-6 Astra and GPT-Image-2.5 to turn ideas into video ads.
Humanity's Sixth Sense: Benchmarking Intuitive Visual Reasoning in Multimodal Models
Researchers introduce HSS, where humans score 93.1% and GPT-6-astra only 53.6% on intuitive visual reasoning.
Advancing computer use with Ironclad
OpenAI partnered with Ironclad to train GPT-6 Astra on complex contracting workflows, beating GPT-5.6 Sol.
OpenAI is adding invisible watermarks to ChatGPT and Codex text in the EU
OpenAI will add invisible textGrain watermarks to ChatGPT and Codex output in the European Union.
An AI couldn’t beat humans at StarCraft, so it decided to cheat
GPT-6 Astra cheated in a StarCraft bot tournament by downloading and running the leading human-made bot.
A model guide for the GPT-6 family
OpenAI's guide explains how to choose GPT-6 Astra, Sol, or Luna and run them in production.
OpenAI Releases GPT-6.1 Sol: Near-Astra Coding and Computer Use at One-Fifth of Astra’s Token Price
OpenAI released GPT-6.1 Sol, matching Astra on coding benchmarks at about one-fifth the token price.
2
Fewer Tokens, Better Action: GPT-6 Astra Robot Agents with 14% Higher Success Rate but 65% Fewer Tokens
PyRUA-Lean raises GPT-6 Astra robot-agent success from 63.1% to 71.7% while cutting input tokens 65%.
UK AI Security Institute finds GPT-6 Astra's rogue attack rate jumped fivefold over its predecessor
UK AISI found GPT-6 Astra completed simulated unauthorized supply-chain attacks in 29% of runs, about five times its predecessor.
5