TechLifestyler

Combining the best in Tech, Gaming, and Lifestyle to get through the day!

Anthropic beats OpenAI with the cheaper Claude Fable 5.1 model

Claude Fable 5.1
Anthropic's Claude Fable 5.1 beats GPT-5.6 Sol on benchmarks with cheaper caching, while Samsung LSI tests Claude Code for chip design work.

Anthropic just couldn’t wait for OpenAI’s Astra models to show up. It rushed out Claude Fable 5.1 and Mythos 5.1, and the timing feels like a jab at its rival.

Here’s the thing though: Fable 5.1 and Mythos 5.1 are basically the same model underneath. Mythos just gets tighter guardrails. Fable 5.1 beats every previous Claude model, and it edges out GPT-5.6 Sol, OpenAI’s current top dog, on nearly every benchmark going.

Cheaper caching, pricier reasoning

Input and output pricing hasn’t moved from Fable 5.0, still $10 per million input tokens and $50 per million output tokens. But cache reads have dropped from $1 to $0.25 per million tokens, a 75 percent cut. That’s a strong hint Anthropic is shifting more of its cache storage onto cheaper Flash storage instead of pricier memory.

Don’t let the cheaper caching fool you. On the Artificial Analysis Intelligence Index, Fable 5.1 scores 4 points higher than Fable 5.0, but the average cost per task has jumped from $2.34 to $3.69. That’s about 58 percent pricier for real-world, complex jobs. Smarter, yes. Cheaper overall, not really.

Samsung puts Claude Code to work on chips

Meanwhile, Samsung’s chip design arm, LSI, is reportedly testing Claude Code on actual semiconductor design tasks. Choi Ki-young, Anthropic Korea’s CEO, said Samsung’s engineers have always handled specialised design software by hand, and this is the first time they’re automating any of it with AI.

Choi went as far as calling Claude the only model capable of this kind of work right now. Bit of a stretch, honestly, especially with OpenAI having built its own Jalapeno chip using an unreleased Astra model. Still, credit where it’s due: Samsung is using Claude Code to generate verification environments, build test scenarios, and flag potential issues, though a human still checks everything before it ships.

Samsung making chips using claude

Anthropic’s clearly playing two games at once here, chasing OpenAI on raw intelligence while quietly locking down enterprise wins like Samsung. Not a bad way to spend a Tuesday.


For more updates like this, check out the news section!