Fable 5 is not the number 1 anymore!? Fugu by Sakana beats it in some cases, according to the company announcement
Summary
Sakana AI's new model Fugu outperforms Fable 5 on LiveCodeBench and Terminal Bench 2.1 by small margins, according to their corporate blog, though results are not yet independently confirmed.
Similar Articles
Sakana Fugu Launches With 93.2 LiveCodeBench Score After Claude Ban (3 minute read)
Sakana AI launched Fugu, a multi-model orchestration API, after Anthropic suspended access to Claude Fable 5 and Mythos 5. Fugu Ultra scored 93.2 on LiveCodeBench, outperforming Fable 5, and starts at $5 per million input tokens.
Fable 5 sits at the top of KernelBench. Jack Clark calls it “the start of a RSI loop”
Fable achieves top rank on KernelBench-Mega by writing a highly efficient CUDA megakernel with an 18.71X speedup, signaling progress toward recursive self-improvement in AI R&D.
Fable 5 vanished in 96 hours and four days later an MIT model took its arena crown
Anthropic's Fable 5 model disappeared after 96 hours due to export controls, and days later, Z.ai open-sourced GLM-5.2 under MIT license, surpassing Fable 5 on the Design Arena. This highlights that the best model is not always the most accessible, shifting focus from benchmarks to availability and licensing.
Fable 5 Is Dead. And Honestly? We Might Be Better Off
US government forced Anthropic to pull its most powerful model, Fable 5, just days after launch. New benchmarks from OpenRouter show that fused panels of budget models can match or exceed Fable 5's performance at half the cost, raising questions about the value of frontier models.
@hqmank: Fable 5 > GPT-5.6 Sol Change my mind.
A user claims that Fable 5 outperforms GPT-5.6, challenging others to disagree.