Agents / OrcaRouter · Qwen base model

Qwen3.8 Flash Next Uncensored

OrcaRouter’s abliterated build of Qwen3.8-Flash-Next, a 177B mixture-of-experts model with tools, thinking and vision. The 98.5 GB IQ4_XS quantization fits a 128 GB Mac or workstation.

Verified sourceSource checked 11/17/2026
LOCALOPEN WEIGHTS
Freshness 72 out of 100. Repository age: 87 days old, Aug 26, 2026. Measures age, not quality.
Popularity 79 out of 100. 673,135 downloads in 30 days; 1,551 lifetime likes. Checked Nov 21, 2026. Measures adoption, not quality.
What do these scores mean?

Two ways to compare models. Neither score measures answer quality, benchmark performance, or whether a model fits your machine.

Freshness · release age

A new release starts at 100. The score halves every 180 days: 100 × 2^(−age in days / 180). When the release date is unavailable, a repository date is a fallback—not proof of when the model was released. Missing dates stay unscored.

Check the date source
Popularity · adoption

Downloads over the past 30 days carry 80% of the score; lifetime likes carry 20%. Both use a logarithmic scale, capped at 10 million downloads and 100,000 likes. If likes are unavailable, downloads carry all the weight. Missing download statistics stay unscored.

round(80 × min(1, log10(1 + downloads) / 7) + 20 × min(1, log10(1 + likes) / 5))

Check the statistics source
Scores as of Nov 21, 2026. Statistics checked Nov 21, 2026. Statistics are marked for refresh after 72 hours.Full scoring method
WORKFLOWS FOR THIS MODEL 1

What do you want to do?

Choose a task. Get the software, files, settings, and steps to run it.

MAKE IT PERSONALChoose your hardware

Open a guide for its software, hardware requirements, and setup steps.

At a glance

Parameters
177B total
Architecture
Qwen4 mixture of experts (qwen4_exp) · 512 experts, 10 routed and 1 shared per token
Context length
262,144 native (Qwen base model) · OrcaRouter’s examples use 8,192
License
Apache 2.0, inherited from Qwen3.8-Flash-Next
Software
llama.cpp
MY HARDWARE

Will it run on your machine?

Save your machine to see a personalized rating and its reasoning.

Add my hardware

Best for

  • Run a local Hermes agent that reads and writes files with your approval.
  • Private coding and reasoning with fewer refusals on a 128 GB Mac Studio or workstation.
  • Ask about images by loading the vision file alongside the model.

Tradeoffs

  • OrcaRouter reports refusals on harmful prompts falling from 64–100% to roughly 0–3.3%, with capability within 2 points of the base model. Those are publisher results for its own files, not for mradermacher’s quantization and not YouRunAI tests.
  • OrcaRouter’s card frames the release for research use and says safety alignment was substantially removed.
  • Needs a llama.cpp build from August 27, 2026 or later; OrcaRouter notes some apps still fail with an unknown-architecture error.
  • OrcaRouter’s own GGUF repository is gated: you must accept its terms on Hugging Face to download from it.
  • A different model from Qwen3.8-Flash-Next · Strata, which is a separate runtime and package on YouRunAI.

Ways to run it

llama.cpp · mradermacher IQ4_XS · Windows, macOS, Linux

Single 98.5 GB file, not gated. Used by the Hermes agent guide on this page.

llama-server -hf mradermacher/Qwen3.8-Flash-Next-Uncensored-GGUF:IQ4_XS --jinja -c 65536 --temp 1.0 --top-p 0.95 --top-k 20 --min-p 0.0
Official instructions

llama.cpp · OrcaRouter Q4_K_M · Windows, macOS, Linux

OrcaRouter’s recommended file, about 110 GB in three parts. Gated: run hf auth login first.

llama-server -hf orcarouter/Qwen3.8-Flash-Next-Uncensored-GGUF:Q4_K_M --jinja -c 8192
Official instructions

What to expect

The pick for 128 GB of memory. Use the 98.5 GB IQ4_XS file in llama.cpp and expect a tight fit.

Explore its uses

UNCENSOREDABLITERATEDAGENTSTOOL CALLINGREASONINGMIXTURE OF EXPERTSVISIONGGUFQWEN3.8
REFERENCE LIBRARY

Sources and files

Primary sources and files behind this model page.

3 SOURCES

Keep exploring