20 August 2026

Anthropic built a stronger Claude model that stays internal only

• Anthropic, the company behind Claude, developed an unreleased model called Model 2 that outperforms all public versions of Claude, according to its August 2026 risk report.
• Model 2 scores 1.5 points higher than Claude Mythos 5 on Anthropic's internal capability scale, a smaller gain than previous public releases showed between versions.
• Anthropic uses Model 2 internally for coding, data generation, and research work, with Claude now writing most code in the company's own production systems.
• The company tested Model 2 less thoroughly than public releases but found no new safety concerns, rating overall misalignment risk as low.
• Anthropic has no current plans to release Model 2 publicly, keeping its strongest model for internal use only.

Reported by The Decoder