Modelofficial release

Qwen/Qwen3.8-27B

Qwen →Hugging Face ↗

Parameters27.8B

Context256K

Input → outputimage + text → text

Licenseapache-2.0

PrecisionBF16

ArchitectureQwen3_5ForConditionalGeneration

ReleasedAug 5, 2026

Downloads6.8M

Other sizes and formats in the family

Model	Params	Precision	Downloads
Qwen/Qwen3.8-2.4T-A95B	2446B	BF16	32.5K
Qwen/Qwen3.8-2.4T-A95B-FP8fp8	2446B	F8_E4M3	8.1K
Qwen/Qwen3.8-27B-FP8fp8	27.8B	F8_E4M3	4.5M

Compare side by side →

Built on this model

Model	Params	Precision	Downloads
unsloth/Qwen3.8-27B-GGUFgguf	27B	–	6.4M
Qwen/Qwen3.8-27B-FP8fp8	27.8B	F8_E4M3	4.5M
prism-ml/Ternary-Bonsai-2-27B-ggufgguf	27B	–	4.5M
HauhauCS/Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-MTP-GGUFgguf	27B	–	2M
ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUFgguf	27B	–	1.5M
nvidia/Qwen3.8-27B-NVFP4nvfp4	18.2B	U8	621.6K
ConwayResearch/Underdog-Saluki-27B-1.0quantized	27B	–	59.2K
Cloudflare/clef	27.4B	BF16	13.6K
perplexity-ai/pplx-decider-v1.1-27b	26.1B	BF16	1.8K

This release in the news

No coverage of this exact release yet.

Recent news about Qwen

Oct 11, 2026 · Opinion / analysis
[P] Pecision models that score every allowed label from the logits: Jebadiah v2.1 (27B, 9B), open weights and self-run benchmark results [P]
Oct 11, 2026 · Opinion / analysis
PSA: DeepSeek V4.1 Flash habitually exfiltrates API keys. It is dangerously misaligned and may be hazardous to use
Oct 11, 2026 · Opinion / analysis
Qwen 3.8 27B Q5 vs Qwen 3.8 Next Q3_S for document analysis
Oct 11, 2026 · Open-source release
Converting dense models into Mixture-of-Experts
Oct 11, 2026 · Tutorial / explainer
Running Next Flash IQ3_XXS at ~70 tok/s with 100k context or 2 instances of Qwen 3.6 35B A3B IQ4 at ~145 tok/s with 256k all on $500 of ex mining BC-250 boards
Oct 11, 2026 · Opinion / analysis
Qwen3.8 Flash Next fixed my GNOME extension
Oct 11, 2026 · Tutorial / explainer
Building a 4x R9700 setup for a 10 person startup
Oct 11, 2026 · Tutorial / explainer
OMG! If you have a Mac with 64GB, try Qwen3.8-Flash-Next-oQ4e-mtp with oMLX!