GLM-5.3-Flash released: 320B params, 1M context, 90% cheaper
baseten · x · 2026-08-27
Z.AI released the GLM-5.3-Flash model, now live on Baseten.
Key Features:
- Architecture & Performance: 320B total parameters (18B active) with a hybrid sparse-and-linear attention architecture; approaches Claude Opus 4.8 on coding and agentic evaluations.
- Multimodal: Native vision support, capable of interpreting interfaces and using browser tools.
- Long Context: Supports up to 1 million token context.
- Cost: 90% cheaper than GLM-5.2, priced at $0.15/1M input tokens and $0.50/1M output tokens.
- License: MIT License.
Related event: Zhipu Open-Sources GLM-5.3-Flash: 320B MoE Multimodal at 1/10th the Cost(48 posts)→
More from Models
- Google says teams across the company are using Gemini 4 Argon on its hardest engineering problems — CurieuxExplorer · 2026-10-15
- Perplexity's new embedding models: index with a 9B model, search with 0.6B — antoine_chaffin · 2026-10-15
- 797GB of n-gram counts in CPU RAM let 1B models match 57% more compute — PangWeiKoh · 2026-10-15
- Leak: Moonshot's Kimi flagship 'x4' done with pre-training, K3.1 in internal testing — lyraxana · 2026-10-15
- Unsloth ships GGUF quants for Nemotron 3.5: Bolt-74B runs in 24GB RAM — danielhanchen · 2026-10-15
- NVIDIA ships two open Nemotron 3.5 Bolt models: a 74B-A9B MoE and a mobile-friendly E4B — HankYeomans · 2026-10-15