Anthropic Launches Claude Opus 4.8 to Improve AI Honesty

Anthropic released its latest flagship AI model, Claude Opus 4.8, on May 28, 2026. Replacing Opus 4.7, the model emphasizes intellectual honesty and a reduced tendency to hallucinate or provide sycophantic answers. Performance testing indicates the model is more likely to challenge user delusions and provide blunt, realistic interventions compared to OpenAI's ChatGPT-5.5 Instant.

Anthropic claims the model outperforms Google's Gemini 3.1 Pro and OpenAI's GPT-5.5 in synthetic benchmarks for computer use, financial analysis, and agentic coding. However, reports from ZDNET suggest the model does not fully surpass GPT-5.5 in certain coding benchmarks. The release introduces dynamic workflows capable of deploying hundreds of subagents and a new Fast Mode that operates 2.5 times faster than standard speed.

Standard pricing remains $5 per million input tokens and $25 per million output tokens. The Fast Mode is priced at $10 per million input tokens and $50 per million output tokens. The model is accessible via the Claude API, claude.ai, major cloud platforms, and Microsoft Foundry. This launch occurs as Anthropic and OpenAI compete for potential public market debuts later in the year.