AI BriefingClaude Opus 4.8 Unveiled
Key point
Anthropic has unveiled Claude Opus 4.8, a new model with significantly improved performance and efficiency.
Details
Anthropic has released Claude Opus 4.8, an improvement over the existing Opus 4.7 with higher benchmark performance and collaboration capabilities. The model's usage pricing remains the same as before.
Along with this update, various new features have been introduced. claude.ai users can now directly control the amount of effort put into a task, and Claude Code has gained a 'dynamic workflows' feature for solving large-scale problems. In addition, fast mode, which is 2.5x faster than the previous model, is now 3x cheaper in cost.
Opus 4.8 shows improved performance across coding, agentic skills, reasoning, and practical knowledge tasks. Key results are as follows:
- Super-Agent benchmark: Achieves an edge by completing all cases at a cost on par with GPT-5.5.
- Legal Agent Benchmark: Recorded industry-leading accuracy.
- Online-Mind2Web: Recorded 84%, demonstrating strong computer use and browser agent capabilities.
- CursorBench: Performs efficient tool calling that surpasses the previous model at all effort levels.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.