METR Finds Recent AI Models, Including GPT-6 Astra, Can Complete Tasks Taking Humans a Full Work Week
AI nonprofit METR said recent AI models, including OpenAI's GPT-6 Astra, can complete tasks that would take a human around a full work week. METR said AI capabilities have continued doubling every 3.6 months.
AI nonprofit METR said Friday that recent AI models, including OpenAI's GPT-6 Astra, can complete tasks that would take a human around a full work week to do, according to a briefing from The Information.
Key findings
- Task length: Recent models, including GPT-6 Astra, are able to complete tasks that would take a human around a full work week, METR said.
- Growth rate: METR said AI capabilities have continued doubling every 3.6 months.
- Method: The finding rests on METR's time horizon measurement.
Benchmark scores, context window, pricing and parameter count for GPT-6 Astra were not part of the report. Neither were the other models METR evaluated, beyond the statement that they are "recent" and include GPT-6 Astra.
What this means
METR's time horizon measurement expresses model capability in terms of how long a task would take a human to complete. On that measure, METR reports that the length of task AI can handle now reaches around a full work week, and that capability has continued to double every 3.6 months. The finding is METR's own, reported by The Information, and it names GPT-6 Astra as one of the models reaching that level.
Related Articles
METR adds open-ended tasks to its AI time horizon plot; models fall short on judgment-heavy work
AI evaluation nonprofit METR has updated its time horizon plot with longer and more open-ended tasks. The report finds some capabilities still improving exponentially, but models fall short of humans on judgment-heavy tasks such as designing AI evaluations.
OpenAI introduces MedChemBench to test AI on early drug discovery tasks
MedChemBench evaluates how well AI models interpret molecular structures, reason from experimental evidence, and select compounds in early drug discovery. It is published on OpenAI's site under the title "Can AI help scientists identify promising drug candidates?"
Sakana AI releases JFinBench, a 321-question Japanese finance benchmark from 2026 certification exams
Sakana AI has released JFinBench, a Japanese financial benchmark of 321 questions drawn from the 2026 editions of six certification exams. Across 12 evaluated LLMs, Kimi K3.1 Pro falls about 10 points behind GPT-6 Astra, and Sakana AI says failures stem mainly from missing knowledge of Japan's systems rather than calculation errors.
Claude Fable 5.2 paints images through Python code, using about 700,000 simulated strokes in one piece
Anthropic's Claude Fable 5.2 produces images by writing Python programs that drive a simulated paint engine, with no external image model. Anthropic reports 64.9 percent on its internal RoboDojo Sim benchmark against 29.8 percent for OpenAI's GPT-6 Astra, though a real-world robot arm test showed limits in speed and fine motor control.
Comments
Loading...