ISTA-DASLab / image-text-to-text Open weights updated 3 days ago

Qwen3.8-Flash-Next-GSQ-RCO-Coder-GGUF

Name: Qwen3.8-Flash-Next-GSQ-RCO-Coder-GGUF
Author: ISTA-DASLab

A 512-expert mixture-of-experts model with half of its experts removed, targeted at code and retaining multimodal capability. 58.4 GB against 354 GB at BF16, an effective 1.89 bits per parameter of the original transformer.

Params

180.0 B

Context

—

Downloads 30d

932 K

Likes

478

Commercial use: allowed apache-2.0 Not gated GGUF View on Hugging Face ↗

Download history

daily snapshots · 8 days

932 K799 K

Oct 16Oct 18Oct 21Oct 23

Can you run it?

Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.

File	Quant	Size	Est. VRAM	Verdict on RTX 4090 · 24 GB
mmproj-Qwen3.8-Flash-Next-BF16.gguf	GGUF	0.9 GB	28.5 GB	❌ Won’t fit
Qwen3.8-Flash-Next-GSQ-RCO-IQ1_M-00001-of-00002.gguf	IQ1_M	29.6 GB	60.1 GB	❌ Won’t fit

Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.

Specifications

Parameters: 180.0 B
Licence: apache-2.0
First seen on the Hub: 2026-09-26
Base model: Qwen3.8-Flash-Next
Training datasets: undisclosed
Added to our catalog: 2026-10-16

Family

Base model and the most-downloaded derivatives in the catalog.

Base Qwen3.8-Flash-Next Qwen 2.7 M

Compare with any image-text-to-text model

Popular comparisons

Qwen3.5-4B 7.6 M Qwen3.5-9B 6.7 M Qwen3.8-Flash-Next-GSQ-RCO-GGUF 6.4 M Qwen2.5-VL-7B-Instruct 5.0 M Qwen3-VL-8B-Instruct 4.8 M Qwen3.8-27B-FP8 4.3 M

Appears in

Fastest-growing VLMs