Claude Mythos-class AI will be available for public in next few weeks, Anthropic confirms
Anthropic has said Mythos-class models will be rolled out to all customers in the coming weeks. The move follows Mythos preview testing that found thousands of serious software vulnerabilities and prompted wider safeguards.
Anthropic on Thursday launched its upgraded Claude Opus 4.8 model, but what caught wider attention was a short paragraph at the end of the announcement. In the release, the company revealed that it is preparing to roll out its Mythos-class model to all customers in the coming weeks. While it remains unclear whether the upcoming release will be Mythos itself or a model similar to Mythos, Anthropic has said that it plans to roll out “Mythos-class” models in the coming weeks.
Mythos is Anthropic’s large language model with advanced cybersecurity capabilities that can reportedly find thousands of bugs in software overnight. The model’s capabilities have emerged as a major cybersecurity concern among executives and world leaders.
Because of its ability to detect hidden cybersecurity vulnerabilities, Mythos is currently available only in preview through Anthropic’s Project Glasswing initiative. It brings together major technology companies including Amazon Web Services, Anthropic, Apple, Broadcom, Cisco, CrowdStrike, and Google to secure critical software systems around the world before the model's wider release.
The company now says it is making rapid progress on those safety protections.
“We’re making swift progress on developing these safeguards and expect to be able to bring Mythos-class models to all our customers in the coming weeks,” Anthropic said in the Claude Opus 4.8 release announcement.
Mythos has already discovered thousands of vulnerabilities
According to Anthropic, Mythos Preview has already identified thousands of high-severity vulnerabilities, including some affecting every major operating system and web browser.
The company claims that most Project Glasswing partners have each discovered hundreds of critical vulnerabilities in their own software using Claude Mythos. Several partners reportedly said their bug-finding rate increased by more than ten times.
Cloudflare said it found around 2,000 bugs across its critical systems, including nearly 400 categorised as high- or critical-severity vulnerabilities.
Anthropic also claimed that Cloudflare’s team found Mythos Preview’s false positive rate to be better than that of human testers.
Indian organisations are also preparing
In India, some of the country’s biggest banks, government departments, and IT companies have quietly started stress-testing critical software systems amid growing concerns around Anthropic’s upcoming Mythos rollout.
At the same time, Anthropic’s newly launched Claude Opus 4.8 model is being positioned as an upgrade focused on stronger agentic capabilities, advanced coding performance, and improved honesty and self-correction.
Anthropic on Thursday launched its upgraded Claude Opus 4.8 model, but what caught wider attention was a short paragraph at the end of the announcement. In the release, the company revealed that it is preparing to roll out its Mythos-class model to all customers in the coming weeks. While it remains unclear whether the upcoming release will be Mythos itself or a model similar to Mythos, Anthropic has said that it plans to roll out “Mythos-class” models in the coming weeks.
Mythos is Anthropic’s large language model with advanced cybersecurity capabilities that can reportedly find thousands of bugs in software overnight. The model’s capabilities have emerged as a major cybersecurity concern among executives and world leaders.
Because of its ability to detect hidden cybersecurity vulnerabilities, Mythos is currently available only in preview through Anthropic’s Project Glasswing initiative. It brings together major technology companies including Amazon Web Services, Anthropic, Apple, Broadcom, Cisco, CrowdStrike, and Google to secure critical software systems around the world before the model's wider release.
The company now says it is making rapid progress on those safety protections.
“We’re making swift progress on developing these safeguards and expect to be able to bring Mythos-class models to all our customers in the coming weeks,” Anthropic said in the Claude Opus 4.8 release announcement.
Mythos has already discovered thousands of vulnerabilities
According to Anthropic, Mythos Preview has already identified thousands of high-severity vulnerabilities, including some affecting every major operating system and web browser.
The company claims that most Project Glasswing partners have each discovered hundreds of critical vulnerabilities in their own software using Claude Mythos. Several partners reportedly said their bug-finding rate increased by more than ten times.
Cloudflare said it found around 2,000 bugs across its critical systems, including nearly 400 categorised as high- or critical-severity vulnerabilities.
Anthropic also claimed that Cloudflare’s team found Mythos Preview’s false positive rate to be better than that of human testers.
Indian organisations are also preparing
In India, some of the country’s biggest banks, government departments, and IT companies have quietly started stress-testing critical software systems amid growing concerns around Anthropic’s upcoming Mythos rollout.
At the same time, Anthropic’s newly launched Claude Opus 4.8 model is being positioned as an upgrade focused on stronger agentic capabilities, advanced coding performance, and improved honesty and self-correction.
