ModelsAug 22, 2026 · 06:496 sources

Ox Alpha Hits 80% on Coding Benchmark, Besting Fable in Stealth Launch

From huggingnews.com

AI;DR

An anonymous AI model called Ox Alpha has been released for free across OpenRouter, Venice, and Cline, featuring multimodal capabilities and a 1 million token context window. The model scored 80% on a DeepSWE coding subset, outperforming Fable, GLM-5.3, and Sol. The Nous Portal claims it can handle 1 quadrillion tokens per day for the model.

Read the sourceSign in to suggest

Sources

+ 5 more
  1. SUPPORT@alexatallah“$ ori [your fav harness] --model stealth/ox-alpha”x.com
  2. SUPPORT@askvenice“This is an experimental release. Ox Alpha is free for all users.”x.com
  3. SUPPORT@cline“Early benchmarks shows marginal improvement over Fable and GPT.”x.com
  4. SUPPORT@nousresearch“We have capacity for 1 quadrillion tokens per day.”x.com
  5. SUPPORT@mtslive“the previous four were all claimed by Chinese labs: Zhipu, Xiaomi, Ant, and Meituan”x.com

One ranked edition, every morning

The stories that matter today. Free, cancel anytime.

Telegram @aidr_today