As competition in the world of artificial intelligence continues unabated, Anthropic has released Claude Fable 5.1 and Mythos 5.1 models in response to OpenAI’s Astra series models. Technology giant Samsung’s LSI division confirmed it has begun testing Anthropic’s Claude Code tool with the aim of optimizing chip design processes. These new model updates, announced with a date of September 1, 2026, are seen as a significant milestone, particularly in AI-powered hardware design processes. Anthropic aims to both raise the performance bar and optimize API costs with Fable 5.1.

Claude Fable 5.1 Model Redefines Performance Standards

The newly announced Claude Fable 5.1 model reveals quite impressive results in benchmark tests. Especially in scientific terminal tests, this model, which performs more than twice as well as the previous model, also surpasses leading competitors in the sector such as GPT-5.6 Sol.

Although the two models have fundamentally similar architectures, Fable 5.1 is equipped with much more robust and reliable system defenses.

On the cost side, Anthropic made a strategic move by reducing cache read prices from $1 per million tokens to $0.25. This step is interpreted as the company transferring KV cache management to Flash storage units more efficiently.

However, it is observed that the unit processing cost increases somewhat in complex real-world scenarios. This shows that the high intelligence capacity provided by the model requires operational budget stability.

Samsung LSI Leverages AI in Chip Design Processes

Statements made by Anthropic Korea CEO Choi Ki-young demonstrate Samsung Electronics’ commitment to integrating artificial intelligence into hardware development processes. Samsung researchers are now trying to automate complex design software processes, previously carried out manually, with the help of Claude Code.

This collaboration serves as a concrete example of how far the application areas of artificial intelligence can expand, especially in the semiconductor world.

Samsung’s work in this area focuses on the code generation and defect detection capabilities provided by the model. Although human control is still necessary for the reliability of the results, it is stated that the speed provided by Claude Code significantly shortens the design process.

Considering OpenAI’s work on its own ASIC chip, Jalapeno, this collaboration between Anthropic and Samsung seems set to further intensify competition in the hardware world.

How long do you think it will take for AI to become fully autonomous in critical engineering areas like chip design? How do you evaluate the performance increase offered by the Claude Fable 5.1 and Samsung’s breakthrough? Share your thoughts in the comments.