Anthropic and OpenAI Say Their Models Are Fast-Tracking AI Research
AI labs are claiming their own models are accelerating the pace of AI research itself. METR estimates Anthropic's benchmark progress has nearly doubled in speed, while OpenAI says its systems are completing more tasks without a human in the loop. Success still swings widely depending on the task.
- Original source: thedefiant.io.
- The source published it on Nov 19, 2026 at 10:34 AM.
- No second outlet has reported it yet. Treat it as a single-source claim.
AI labs are now making the case that their own models are speeding up the work of AI research. METR estimates that Anthropic's benchmark progress has nearly doubled in pace, a figure the labs are treating as evidence their systems are moving faster through research tasks.
OpenAI reports higher estimated task-success rates without human intervention — meaning the models are getting more done on their own. The company notes that improvement swings widely depending on the task, so the gains aren't evenly spread across everything the models are asked to do.
TopicsAnthropic · OpenAI · METR
Source: thedefiant.io
Written by AI from verified reporting.
More stories
- Tech & AI reuters.com Caught by @mimi💕💕 0 viewsNorth Korea rolls out its first AI chatbot, but politics is off-limits
- Tech & AI cryptobriefing.com 0 viewsMicron tumbles 40% off record high as AI spending doubts rattle chips
- Tech & AI cryptobriefing.com 3 sources 1 viewAnthropic's $107.3B record IPO hauls Bitcoin back toward $64,000
- Tech & AI finance.yahoo.com 1 viewNvidia, Intel Snap 5-Day Rout as Trump Floats Blackwell China Sales

