METR updates its AI task-time curve: coding capabilities still scaling exponentially
11/20/2026 — 11/21, 04:17·2 sources (all)·2 reports (all)
Key takeaways
- AI evaluation nonprofit METR has released a new report updating its widely cited AI task-time curve, finding that some capabilities still fo…
Background & analysis
On November 20, 2026, at 16:00, the AI evaluation nonprofit METR released a report arguing that model capabilities tied to AI R&D are improving rapidly while becoming increasingly difficult to measure. The report's main body is organized into sections on background, methodology, results, robustness checks, and discussion. Its appendices cover difficulty estimation and baselining, time horizon diagnostics, the methodology behind MCI modelling, limitations of ITRE MCI robustness checks, estimating lifecycle cost for tasks, data integrity and reward hacking, changes in capability growth rates, inference scaling, swarm experiments, Al Zimmermann's Programming Contests, budget NanoGPTM, making METR tasks, and Earthborne Rangers and continual learning.
On November 21, 2026, at 02:15, MIT Technology Review · AI reported that METR had put out a new report updating the organization's widely cited time horizon plot. The report, as described in that coverage, finds that some AI capabilities are still improving along an exponential trajectory, while also indicating that models do not perform as well on tasks requiring judgment as they do on coding tasks. The two accounts agree that METR published a report and that AI capabilities continue to progress; the first emphasizes that such gains are growing harder to measure, while the second highlights the gap between coding and judgment-based performance. As of now, the story stands at the point where METR's report has been published and is being circulated through external coverage.
AI-generated from 2 reports · updated 1 day ago
Latest turnThe AI evaluation nonprofit METR has published a new report updating its widely cited time horizon plot, finding that some AI capabilities are still improving along an exponential trajectory. The report also indicates that models remain weaker at tasks requiring judgment than at coding.
- Heat index
- 69
- All sources
- 2
- All reports
- 2
- First seen
- 2 days ago
Reports on this story headlines open the original
The AI evaluation nonprofit METR has published a new report updating its widely cited time horizon plot, finding that some AI capabilities are still improving along an exponential trajectory. The report also indicates that models remain weaker at tasks requiring judgment than at coding.
MIT Technology Review · AIAI score 56
METR published a report finding that AI R&D-relevant model capabilities are improving rapidly while becoming harder to measure. The report covers methodology, results and robustness checks, and discusses task difficulty estimation, time-horizon diagnostics and changes in capability growth rates.
METRFirst-partyAI score 53
Other stories people are talking about
- Heat index 1143RisingAnthropic Reaches $1.6 Trillion Market Value in First Trading Week After IPO21 sources
- Heat index 1011Anthropic Mathematician Says Claude Formalized the Generalized Riemann Hypothesis21 sources
- Heat index 622Anthropic IPO total raise reaches $93.5 billion after greenshoe exercise20 sources
- Heat index 596RisingThinking Machines Lab Open-Sources Inkling 2 Mini with 32GB Model Card Recommendation10 sources
- Heat index 490RisingSEC probes Nvidia accounting practices as shares post seven-week slide9 sources
How is heat calculated?About the methodHide
Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.
This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.
- Surge
- Discussion rising fast
- New
- First report within 6 hours
- Rising
- Still gathering discussion

