kindlingai/glm-5.3-exlr8-k2-k3-k4
just landedmodel
EXLR8 is a new quantization format. Mixture-of-experts (256 experts, 8 per token). Context: 1M tokens (config). License: glm-5.3. Vendor-reported: "GLM-5.3's top-8 routing turns small numeric differences into different experts: two BF16 runs that differ only in their batching score mean 0.11 (median 0.004) against each other on 32,768 positions of the held-out text."
- Vendor-reported: GLM-5.3's top-8 routing turns small numeric differences into different experts: two BF16 runs that differ only in their batching score mean 0.11 (median 0.004) against each other on 32,768 positions of the held-out text.
- Org
- kindlingai
- Released
- 2026-10-08
- Status
- just landed
- License
- glm-5.3 (restricted)
- Architecture
- MoE · 256 experts, 8/token · GlmMoeDsaForCausalLM
- Context
- 1,048,576 tokens
- Task
- text-generation · text
- Signal
- 23 HF likes · 983682 downloads
Links
First published here 2026-10-08 · updated 2026-10-22 · markdown · JSON