Ox-alpha solves coding task others failed; GLM-5V matches token count
eyishazyer · x · 2026-08-21
Independent tests revealed that ox-alpha single-handedly solved a coding task where GLM-5.3, GPT-5.6-sol, and Grok 4.6 all failed. Additionally, forensic decoding of a 2-second test video showed that ox-alpha spent exactly 296 tokens, with GLM-5V-Turbo exhibiting identical token usage across four separate test videos.
Related event: Mystery Model ox-alpha Fingerprinted as Zhipu's GLM(2 posts)→
More from Models
- poolside researchers on turning tens of thousands of data-mix experiments into frontier models — arena · 2026-08-21
- V4-Flash-Vision tested: 2x faster, cheaper and better than V4-Flash-0731 — cedric_chee · 2026-08-21
- Kimi K3.1 quietly testing on Code Arena; Ox Alpha said to be Zhipu's GLM 5.3 Flash — i_dg23 · 2026-08-21
- Researcher: Agentic coding is the new text summarization—any model gets good results — mishig25 · 2026-08-21
- ThursdAI weekly: Qwen 27B and GLM 5.3 beat GPTs; OpenAI pauses RL for security — thursdai_pod · 2026-08-21
- Ox Alpha's stunning fluid sim 'one-shot' appears to be a copy of an existing GitHub repo — scaling01 · 2026-08-21