Anthropic Unveils Claude Opus 4.8, a New, More Powerful Model

May 28, 2026 - 23:16

Anthropic has quietly released its latest artificial intelligence model, Claude Opus 4.8, marking a significant step forward in the company's ongoing competition with rivals like OpenAI and Google. The new system, called Opus 4.8, tops industry standard benchmarks in areas related to computer programming, showing notable gains in code generation, debugging, and complex reasoning tasks.

According to internal testing data shared by the company, Opus 4.8 outperforms its predecessor, Claude 3 Opus, on nearly every major evaluation metric. The model scored higher on the HumanEval coding benchmark and demonstrated improved performance on math and logic puzzles. Anthropic claims the upgrade also reduces the rate of "hallucinations" -- instances where the AI confidently produces incorrect information.

The release comes as businesses increasingly rely on large language models for software development and data analysis. Claude Opus 4.8 is designed to handle longer documents and maintain context over extended conversations, making it more useful for enterprise clients working on large codebases or complex technical documentation.

Anthropic has not disclosed the exact number of parameters in Opus 4.8, but the company says the model was trained using a refined version of its "constitutional AI" approach, which aims to align the system's behavior with human values and safety guidelines. The model is available now through Anthropic's API and the Claude chat interface, with pricing remaining the same as the previous Opus tier.