Qwen3.8-2.4T-A95B FP8 quantized weights published
An FP8 quantized version of the Qwen3.8-2.4T-A95B MoE text-generation model was published on Hugging Face under the Qwen organisation. The base model is the full-precision Qwen3.8-2.4T-A95B, and this is a quantized variant. The repository listing contains no benchmark results or capability claims.
PUBLISHED2026-08-08
OBSERVED2026-08-21
AGE18d
SOURCES1
- FP8 quantized variant of
Qwen/Qwen3.8-2.4T-A95B - Pipeline: text-generation via transformers
- Tags: fp8, qwen3_5_moe_text, conversational, endpoints_compatible
COMMUNITY
No curated reactions recorded for this event. Facts and takes are kept in separate layers — community context is added by hand, never blended into the record above.