Rumor: GLM 5.5 to Hit 2.5T-3.5T Parameters With Looped Transformer Architecture
scaling01 · x · 2026-10-22
Rumors suggest Zhipu's next model, GLM 5.5, will have 2.5T-3.5T parameters along with significant architecture changes. In its August HKEX interim filing, Zhipu said it had validated the idea of Looped Transformers in preliminary experiments.
Commentator @scaling01 predicts a "cambrian explosion of dirty loopers" next year, with many models adopting looped architectures. All of this remains unconfirmed rumor and personal speculation.
Related event: Rumor: GLM-5.5 to Pack 2.5-3.5T Parameters with Looped Transformers(2 posts)→
More from Models
- Perplexity's JOLT matches GRPO with 6.5x fewer training tokens, JOLT+ wins on all five benchmarks — perplexity_ai · 2026-10-22
- Recursive AI ships Flint, a 40B-A2.5B MoE claimed to match Qwen3.6-27B, weights on Hugging Face — ChengleiSi · 2026-10-22
- DeepSeek ships V4.1-Flash-1017, now the #1 open model on RSI-Inference Bench — scaling01 · 2026-10-22
- OpenAI's Cybersecurity Guardrails Block a Dev From Auditing His Own Patched Server — Next_Technology6361 · 2026-10-22
- TPN-006 hits Hugging Face: 7.7GB compressed model beats its 17GB Mistral teacher — markjeffrey · 2026-10-22
- Kimi K3.1's latest checkpoint draws 49k-token SVG portrait zero-shot in 22 minutes — lyraxana · 2026-10-22