← sources

Probable GLM model

source: reddit (reddit:LocalLLaMA)published 2026-11-15 15:21ingested 2026-11-16 02:38

retained text

https://preview.redd.it/k7ife2j2ms1i1.png?width=591&format=png&auto=webp&s=e8dd29b327ac46e7634a5faf48454ec1beb7aa0b

Because of Theo gg tweet of all people,  
I Launch QFN DeepResearch swarm agent to find anything about this.

**THREE** model names, blurred out, sitting at 59.4% / 57.1% / 54.0%.

1. The blurred names are most likely **\[GLM-Nano or GLM-Name or GLM-Next** (suggestion from Silver-Theme7151)\] Pro, Med, and Small. One community experiment reconstructed the suspected text, applied Gaussian blur (σx ≈ 6.9px), and achieved an RMSE of just 1.254/255 against the original.
2. AICodeKing track record.  Someone compared his previous GLM statements against official release dates: GLM-4.6 (same day), GLM-5 (1 day), GLM-5.1 (11 days), GLM-5.2 (3 days), and GLM-5.3 (4 days).
3. Antrhopic timing. The leaderboard surfaced on November 12 vs AICodeKing timing, while Anthropic IPO on Nov 18. If [Z.ai](http://Z.ai) releases GLM-Nano this week, the timing would make for quite a coincidence

[https://linux.do/t/topic/3166447](https://linux.do/t/topic/3166447)

Another strong signal is HF activities, i know it is not bullet proof, just another signal, 3-4 days compare to last 27 days

https://preview.redd.it/oa0ix3vpls1i1.png?width=767&format=png&auto=webp&s=cbb231665af7d35257fb4d8f24617a2d70bf107a

https://preview.redd.it/1od88gcens1i1.png?width=1600&format=png&auto=webp&s=beb3d5f055dc070dcb92265ed1b91a42cda34e7a

I use LLM to fix my grammar.  
Since it is speculation and not actual open model this post has a high probability to be removed, so yeah

document fetched 2026-11-16 00:37 · canonical https://reddit.com/r/LocalLLaMA/comments/1xwy3ka/probable_glm_model