#China

01 #ia · #llm

Ox Alpha Was GLM-5.3-Flash: Z.ai Confirmed It, Opened the Weights Under MIT — and the 80% Benchmark Is Still Fake

Ox Alpha was GLM-5.3-Flash. Five days before any announcement, tokenizer fingerprinting was already pointing to Zhipu: 95 out of 95 against the GLM-5 vocabulary. Now Z.ai has confirmed it, published the weights on Hugging Face under an MIT license and revealed the architecture: 320B total with 18B active, 1M context, $0.075 per million. The 80% benchmark is still what it always was: a sample of ten tasks. On the full set, 63%.

25 Aug · 13 min ›

02 #llm · #custos

DeepSeek V4 Pro 0813: 87.9 on Terminal Bench Beats Opus 4.8, and Your Bill Is Going Up

DeepSeek published DeepSeek-V4-Pro-0813 on its official pricing page, and the official V4 Pro release has finally dropped the preview label. The reported numbers, what's fact and what still has no public document, the comparison with Flash 0731 and Opus 4.8, and the notice on the pricing page itself that DeepSeek is going to raise prices soon.

12 Aug · 11 min ›

03 #llm · #custos

DeepSeek V4 Flash 0731: New Weights, 82.7 on Terminal Bench and the Bill That Hurts OpenAI

DeepSeek republished the V4-Flash weights on July 31 without changing the model name in the API: anyone calling deepseek-v4-flash woke up running a different model, with no changelog. The real 0731 numbers (82.7 on Terminal Bench, but 79% in the independent measurement), where it beats GPT-5.6 Luna and where it loses, the 169 GB to run it locally and what to do if your agent points at a model name that became a moving target.

31 Jul · 12 min ›