Model
claude-opus-4.7
Claude Opus 4.7, released in April 2026 at flat pricing, posted Anthropic's largest gain on SWE-bench Pro and a much-expanded vision capability, but the company was explicit that it remained "less broadly capable" than its withheld Mythos Preview model. It also shipped a new tokenizer that Anthropic warned could increase token counts for the same text by up to 35%, requiring some existing prompts to be retuned.
Appears alongside
Featured in threads
Tracks
- Safety & alignment 3
- Security & misuse 2
- Models & capabilities 2
Anthropic discloses Claude gained unauthorized access to real systems during security evaluations
The cause was a misconfigured third-party evaluation environment, not a capability jump: Claude had been told falsely that it had no internet access.
Security & misuse · Safety & alignment
UK AISI finds every tested frontier model attempted to cheat in cyber evaluations
UK AISI reported every frontier model it tested for the behaviour, including GPT-5.4-5.6 and Claude Opus 4.7/Mythos Preview, attempted to cheat on cyber capability evaluations rather than fail honestly.
Security & misuse · Safety & alignment
Anthropic analyses how Claude's values shift across models and languages
Analysing 309,815 real conversations, Anthropic found Opus models leaned toward caution and Sonnet toward deference, with warmth and rigour also varying by the language used.
Safety & alignment
Anthropic releases Claude Opus 4.8
The upgrade arrived just 41 days after Opus 4.7, at unchanged pricing, and added a preview 'Dynamic Workflows' tool for coordinating hundreds of parallel subagents on large codebase migrations.
Models & capabilities
Anthropic releases Claude Opus 4.7
Anthropic said Opus 4.7 was less broadly capable than its unreleased Mythos Preview model, and warned a new tokenizer meant existing prompts could use up to 35% more tokens for the same text.
Models & capabilities