Also: How to learn cloud code for free with Anthropic’s AI courses \u2013 one only took me 20 minutes<\/strong><\/p>\nAISI’s updated testing also exemplifies that capability improvements are not limited to individual model releases, but can also occur within versions of a single model. <\/p>\n
rapidly growing cyber threat <\/h2>\n
AISI noted that AI models are rapidly advancing in their ability to handle cyber tasks with serious implications for cybersecurity, especially given the ability of Mythos to detect software vulnerabilities. <\/p>\n
\u201cIn February 2026, we internally estimated that the duration of cyber tasks that could be completed by AI models would double every 4.7 months by the end of 2024 \u2013 already an acceleration from our November 2025 8-month estimate,\u201d the blog authors wrote. “Since then, AISI reported on two new models, Cloud Mythos Preview and (OpenAI’s) GPT-5.5, both of which significantly exceed doubling rate trends.” <\/p>\n
Also: Third major Linux kernel flaw found in two weeks \u2013 thanks to AI<\/strong><\/p>\nThe authors said it is unclear whether this trend will persist or whether these findings indicate a permanent increase. The Mythos and GPT-5.5 models may be notable breaks from the overall pattern of development. <\/p>\n
Nevertheless, AISI clarified that there were many unknowns that could not be determined in its testing. The tests limited the tasks to 2.5 million tokens, allowing researchers to better compare performance results over time. \u201cThis naturally demonstrates what marginal models can do,\u201d he wrote. <\/p>\n
The blog continued, “Mythos Preview and GPT-5.5 have large upper bound error bars due to our narrow cyber suite’s near 100% success rate on the longest tasks, even with the 2.5M token limit.” “Our experiments are not long enough to determine how rapidly the model’s reliability will deteriorate over high task periods. This puts some of the newest models at the measurement limits of our narrow test suite.”<\/p>\n
Also: I put GPT-5.5 through a 10-round test: it scored 93\/100, losing points only due to excitement<\/strong><\/p>\nWhile this makes it harder to measure the model’s failure point, it also means that the model’s success rate on these tasks would be much higher than without the token limit \u2013 so high, in fact, that “it becomes impossible to calculate the time horizon.” Models with greater token access and complex agent infrastructure will be more efficient. <\/p>\n
The blog states, “The 2.5M token limit is relatively low \u2013 in our Cyber \u200b\u200bRange experiment we use up to 100M tokens and find that beyond that budget there is still potential for performance improvement, especially for recent models, which disproportionately benefit from the higher token limit.” <\/p>\n<\/div>\n