---
format: "aidr-story-markdown/v1"
id: "fd7125a01efa24e7351619faa70357ad9c7a84d4aa6ef6eb811aa90f26324694"
canonical_url: "https://aidr.today/fd7125a0?lang=en"
title: "Anthropic Blocks Public Release of Model 2 Following Risk Upgrade to Low"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-08-15T16:44:56.000Z"
category: "Regulation"
topics: ["anthropic","safety","risk"]
source_urls: ["https://huggingnews.com/ai/anthropic-blocks-public-release-of-model-2-following-risk-upgrade-to-low-e8949740","https://x.com/IntCyberDigest/status/2088611503695946021","https://x.com/jun_song/status/2088636373339365708","https://x.com/axios/status/2088344932637438429","https://x.com/GerritD/status/2088417823274262793","https://x.com/GerritD/status/2088420656740495867","https://x.com/GerritD/status/2088419720886456679"]
summary: "Anthropic's August 14 risk report identifies an internal system called \"Model 2\" that is more capable than its Mythos 5 model but will not be released to the public. The company raised its misalignment risk rating from \"very low\" to \"low\" citing uncertainty following recent cyber evaluation incidents. The stronger internal model provides a \"noticeable improvement\" for coding, data generation and agentic tasks. This decision to restrict access aligns with recurring statements from company leadership that certain frontier systems are too dangerous to release to the public."
---

# Anthropic Blocks Public Release of Model 2 Following Risk Upgrade to Low

> [Open the canonical story](<https://aidr.today/fd7125a0?lang=en>)

**Published:** 2026-08-15T16:44:56.000Z
**Category:** Regulation
**Topics:** anthropic, safety, risk

## Summary

Anthropic's August 14 risk report identifies an internal system called "Model 2" that is more capable than its Mythos 5 model but will not be released to the public\. The company raised its misalignment risk rating from "very low" to "low" citing uncertainty following recent cyber evaluation incidents\. The stronger internal model provides a "noticeable improvement" for coding, data generation and agentic tasks\. This decision to restrict access aligns with recurring statements from company leadership that certain frontier systems are too dangerous to release to the public\.

## Sources

- [Story source](<https://huggingnews.com/ai/anthropic-blocks-public-release-of-model-2-following-risk-upgrade-to-low-e8949740>)
- [Story source](<https://x.com/IntCyberDigest/status/2088611503695946021>)
- [Supporting source](<https://x.com/jun_song/status/2088636373339365708>)
- [Supporting source](<https://x.com/axios/status/2088344932637438429>)
- [Supporting source](<https://x.com/GerritD/status/2088417823274262793>)
- [Supporting source](<https://x.com/GerritD/status/2088420656740495867>)
- [Supporting source](<https://x.com/GerritD/status/2088419720886456679>)