Anthropic’s Claude Oceanus-v1-p Opens to Red Team Testing, but Distribution is Compromised

A next-generation Anthropic model has surfaced in restricted testing channels, but early distribution was already compromised before the evaluation formally began.

References to claude-oceanus-v1-p began circulating among researchers on June 3, 2026, after the model identifier appeared inside Anthropic’s Claude Console and surfaced through unauthorized API proxy services.

The sightings immediately triggered speculation that Anthropic was advancing toward a broader rollout of a successor to the Claude Mythos line, with red team evaluators reporting access to the new model beginning that same day.

The controlled evaluation was short-lived. Within hours of the model reaching validated red teamers, reports emerged that an unidentified actor had allegedly resold API access to claude-oceanus-v1-p through a Chinese-based proxy service at a premium rate of $16 per million input tokens, a figure significantly above Anthropic’s standard enterprise pricing tiers.

Anthropic’s history with unauthorized proxy abuse is well-documented; earlier in 2026, the company accused Chinese AI labs including DeepSeek, Moonshot AI, and MiniMax of using approximately 24,000 fake accounts to run over 16 million interactions with Claude models through proxy channels.

In response to the Oceanus resale incident, Anthropic reportedly paused model access for the broader red team cohort pending an internal investigation.

Claude Oceanus-v1-p is understood to build directly upon the Claude Mythos Preview foundation, which launched in April 2026 and demonstrated an alarming capability profile for the cybersecurity community.

Mythos Preview, operating under Anthropic’s restricted research track, was assessed by the company’s Frontier Red Team as capable of identifying and exploiting zero-day vulnerabilities across every major operating system and web browser, with Glasswing partners collectively uncovering over 10,000 high or critical-severity vulnerabilities since the program’s inception.

The Turing Institute further noted that Mythos’ red team found vulnerabilities with a recovery rate exceeding 99% across disclosed test cases.

The Oceanus red team evaluation comes on the heels of Anthropic’s June 2 expansion of Project Glasswing its restricted AI cyberdefense initiative to approximately 150 new organizations spanning more than 15 countries, including India, France, Germany, South Korea, and Australia.

The expanded group now includes important infrastructure sectors like power, water, healthcare, and communications. These sectors were not part of the program when it first launched with a focus on Big Tech.

Anthropic stated that a successful cyberattack on most new partner organizations could affect in excess of 100 million people.

Anthropic has stated candidly that Mythos-level capabilities and, by extension, Oceanus-v1-p, will not be cleared for general public release until the company develops “highly robust safeguards to prevent misuse,” acknowledging that such safeguards do not yet exist in the industry.

Follow us on Google News, LinkedIn, and X to Get More Instant Updates.