METR Warns AI Capability Growth Is Outpacing Safety Testing
METR's latest quarterly capability update finds frontier models now outperform experienced human professionals on most well-specified coding and computational tasks, with the task length they can handle doubling every 3.6 months; a naive extrapolation reaches 50% success on 200-hour tasks by July 2027. The organization warns that current public data cannot rule out a drastic acceleration in AI research, potentially leaving humans unable to meaningfully assess and react to catastrophic risks.
Related Entities
Full Report
AI is getting better faster than experts can reliably test whether new systems are safe, METR warns
AI models are rapidly improving at AI research tasks, and they're getting harder to measure. According to the AI research organization METR, current public data isn't enough to rule out a drastic acceleration in AI research over the next few years. That could leave humans unable "to meaningfully understand, assess, and react to catastrophic risks posed by new AI systems," the organization warns.
On clearly defined coding and computational tasks with direct feedback, the models already beat experienced human professionals most of the time, METR found, though they haven't yet outperformed large groups of top experts. For tasks that take humans longer, the task length the models can handle doubles every 3.6 months. A naive extrapolation of METR's data puts them at a 50 percent success rate on the longest, 200-hour tasks by July 2027.

METR's analysis shows the steady rise in capabilities of leading AI models up through Opus 5.5 and GPT-6 Astra. | Image: METR
The systems still struggle with open-ended questions that lack immediate feedback and require real judgment, as well as with situations that are hard to recreate for training, according to METR. The organization also warns that common, easy-to-run tests may overestimate how well the models perform in real-world use, since developers train them in very similar test environments.
THE DECODER+
- Every Plus article and the full archive
- No ad banners
- Frontier Radar: 6 deep dives a year
What is included → First month −50% Monthly $4.99 for the first month Then $9.99 a month. Cancel any time. Subscribe monthly Best value Yearly $8.25 a month $99 once a year. Subscribe yearly
Already a subscriber? Sign in
Source: [METR](https://metr.org/blog/2026-11-20-q4-2026-capability-update/) [Comment](https://the-decoder.com/subscription) [Copy link](https://the-decoder.com/ai-is-getting-better-faster-than-experts-can-reliably-test-whether-new-systems-are-safe-metr-warns/) [Share](https://the-decoder.com/ai-is-getting-better-faster-than-experts-can-reliably-test-whether-new-systems-are-safe-metr-warns/) Earlier coverage [Nov 22, 2026 **Security researchers show how cheap AI inference providers steal keys and snoop on users**](https://the-decoder.com/security-researchers-show-how-cheap-ai-inference-providers-steal-keys-and-snoop-on-users/) [Nov 18, 2026 **OpenAI opens its AI systems to four independent on-site auditors for the first time**](https://the-decoder.com/openai-opens-its-ai-systems-to-four-independent-on-site-auditors-for-the-first-time/)