Chinese AI Models Scale Up: GLM 5.3 Flash Trains on 30T Tokens
teortaxesTex · x · 2026-08-28
Leaked details reveal progress in Chinese large-scale pre-training: GLM 5.3 Flash is an 18B active MoE architecture pretrained on 30T tokens, while OpenPangu 2.0 Pro boasts 505B total parameters (18B active) trained on Ascend chips. Other models like MooreThreads (236B MoE) and LongCat 2.0 (1600B total) suggest compute barriers are lowering.
More from Infra
- SK Hynix ties new US plant to looser chip equipment exports to China, sources say — pstAsiatech · 2026-10-15
- Perplexity's new embedding models: index with a 9B model, search with 0.6B — antoine_chaffin · 2026-10-15
- Unsloth ships GGUF quants for Nemotron 3.5: Bolt-74B runs in 24GB RAM — danielhanchen · 2026-10-15
- vLLM-Omni ships Day-0 serving for FLUX 3 Image Klein with BF16, FP8, NVFP4 support — vllm_project · 2026-10-15
- NVIDIA's Nemotron 3.5 Super VL lands training and inference support on Applied Compute — ypatil125 · 2026-10-15
- CoreWeave bond yield tops 16% as loans outlast customer contracts by ~2 years — tengyanAI · 2026-10-15