[whatmodel.sh](../../README.md) / Qwen3.8-Flash-Next
# Qwen3.8-Flash-Next
Alibaba · released 2026-08 · 180B params (6B active) · qwen-community-1.0
**score 73.4** ±0.9 · #47 of 732 · from 14 benchmark results across 5 sources
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks Qwen3.8-Flash-Next among every model on that task or test, worst to best.
task vs the others rank counts tests
agent tasks ▁▁▃▆█▄▅▅●▄▂▁ #51 of 313 34% 4
coding ▁▁▁▃▅▇█▇▅▄●▁ #23 of 469 37% 6
vision & documents ▁▁▂▂▁▃▄▂▆▇█● #13 of 153* 10% 2
agent tasks · #51 of 313 · 34% of its score
Beats [Gemini 3.1 Pro (Preview)](gemini-3-1-pro.md) here, though they rank higher overall. Trails [Claude Sonnet 4.6](claude-sonnet-4-6.md), [GLM-5.2](glm-5-2.md), [GLM-5.3-Flash](glm-5-3-flash.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
Arena Agent (October 2026 aggrega… -0.0402 IPS ▁▁▁▃●▂█▁▂▂▁▁ beats 17/59 avg -0.379 IPS 23% sharp
SWE-bench Multilingual (maker) 81.0% ▁··▁▃▁▅██●▂▁ beats 56/65 hard · avg 29% 5.2% typical
SWE-bench Pro public set, Qwen-re… 62.5% ▂·····▂·▂█▃● beats 11/12 hard · avg 21% 3.9% at its level
DeepSWE v1.1 (resolved) (maker) 58.7% ▂····▂▁▁●███ beats 7/31 hard · avg 16% 1.2% at its level · noisy
coding · #23 of 469 · 37% of its score
Beats [GPT-5.6 Sol](gpt-5-6-sol.md), [GPT-5.5](gpt-5-5.md), [MiMo-V2.6-Pro](mimo-v2-6-pro.md) here, though they rank higher overall. Trails [Grok 4.7](grok-4-7.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
Arena WebDev (Code Arena) 1628.3 ▁▁▃▂▅█▃▆▅●▁▂ beats 108/129 avg 981 16% sharp
Luau Leetcode fully solved (Torpe… 76.4% ▁▁▁▁·▁▂▃▅●█▇ beats 26/49 hard · avg 17% 6.7% sharp · at its level
SWE-bench Multilingual (maker) 81.0% ▁··▁▃▁▅██●▂▁ beats 56/65 hard · avg 29% 5.2% typical
SWE-bench Pro public set, Qwen-re… 62.5% ▂·····▂·▂█▃● beats 11/12 hard · avg 21% 3.9% at its level
LiveCodeBench v6 (window unspecif… 91.9% ▃▅█▅▅▃▅···▃● beats 15/16 hard · avg 35% 3.7% typical
DeepSWE v1.1 (resolved) (maker) 58.7% ▂····▂▁▁●███ beats 7/31 hard · avg 16% 1.2% at its level · noisy
vision & documents · #13 of 153 · 10% of its score
Beats [Claude Fable 5.1](claude-fable-5-1.md), [Claude Sonnet 5.5](claude-sonnet-5-5.md), [GPT-6 Sol](gpt-6-sol.md) here, though they rank higher overall. Trails [Muse Spark 1.2](muse-spark-1-2.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
CharXiv (reasoning questions) (ma… 84.6% ▄▂▄▆▄▄▆··▄●▂ beats 20/21 medium · avg 38% 9.4% typical
PureDocBench document parsing (Av… 68.5% ▂·▁·▂▂●█·▁▂▂ beats 10/22 medium · avg 62% 0.2% blunt · at its level
other tests (5)
test result vs the others in its run difficulty counts why
GPQA Diamond (maker) 91.7% ▁▁▁▁▂▄▅▅██●▄ beats 294/322 hard · avg 40% 12% sharp
Humanity's Last Exam (maker) 35.9% ▁▁▁▄▆▆▇█●▆▁▃ beats 87/111 very hard · avg 5.2% 6.2% at its level
BRIDGE clinical text, zero-shot a… 46.9% ▁▁▁▁▁▂▂▃▄▆█● beats 97/97 hard · avg 31% 5.8% blunt · at its level
IFBench (aggregation unspecified)… 81.3% ▂▆▂▄▄▄▂▄▄▆█● beats 83/86 medium · avg 43% 3.5% typical
RobloxQA v2.0 multiple choice (To… 90.7% ▁▁▃▁●▄▃▃▆▂▂▂ beats 17/49 easy · avg 77% 2.3% maxed out
source results share of the score
Qwen (submitted / maker-reported) 8 45% ▇▇▇▇▇
LMArena leaderboard dataset 2 40% ▇▇▇▇▇
Roblox LLM Leaderboard 2 9% ▇
BRIDGE Medical Leaderboard 1 6% ▇
How Far Is Document Parsing from Solved… 1 0% ▇
prior: pull toward the middle 0% ▇
weighted average = 73.4
Licences: LMArena leaderboard dataset ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · Roblox LLM Leaderboard ([MIT](../../licenses/roblox-llm-leaderboard-MIT.txt)) · BRIDGE Medical Leaderboard ([MIT](../../licenses/bridge-medical-leaderboard-MIT.txt)) · How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/))
## How to run it
### llama.cpp (recommended)
1. Install llama.cpp. Homebrew and winget installs name the commands llama-cli and llama-server instead. [Other ways to install](https://github.com/ggml-org/llama.cpp/blob/master/docs/install.md)
```sh
# macOS / Linux
curl -LsSf https://llama.app/install.sh | sh
# Windows
irm https://llama.app/install.ps1 | iex
```
2. Run Qwen3.8-Flash-Next (ud-q2_k_xl, needs ~54.5 GB, from [huggingface.co/unsloth](https://huggingface.co/unsloth/Qwen3.8-Flash-Next-GGUF/tree/main))
```sh
llama cli -hf unsloth/Qwen3.8-Flash-Next-GGUF:UD-Q2_K_XL
```
3. Or start a local server and open the address it prints (http://localhost:9931)
```sh
llama serve -hf unsloth/Qwen3.8-Flash-Next-GGUF:UD-Q2_K_XL
```
### LM Studio
1. Install LM Studio. Open the app once before using the lms command. [Download it (macOS / Windows / Linux)](https://lmstudio.ai/download)
2. Run Qwen3.8-Flash-Next (ud-q2_k_xl, needs ~54.5 GB, from [huggingface.co/unsloth](https://huggingface.co/unsloth/Qwen3.8-Flash-Next-GGUF/tree/main))
```sh
lms get https://huggingface.co/unsloth/Qwen3.8-Flash-Next-GGUF@UD-Q2_K_XL -y
```
3. Then run lms ls to see its name, and lms chat followed by that name
### Ollama
1. Install Ollama. [Download it (macOS / Windows)](https://ollama.com/download)
```sh
# Linux
curl -fsSL https://ollama.com/install.sh | sh
```
2. Run Qwen3.8-Flash-Next (Q4_K_M, needs ~108.5 GB, from [ollama.com](https://ollama.com/library/qwen3.8-flash-next/tags))
```sh
ollama run qwen3.8-flash-next:125b-a6b-q4_K_M
```
### Strata
1. Install Strata. Runs Qwen3.8-Flash-Next split across your graphics card, RAM and SSD: an NVIDIA or AMD card with 12 GB or more, Windows or Linux.
```sh
# Windows / Linux
git clone https://github.com/Niko1221/Strata.git && cd Strata
```
2. Run Qwen3.8-Flash-Next (UD-IQ4_XS, needs 12 GB VRAM + 80 GB RAM, 93.7 GB download, NVIDIA or AMD card only, from [github.com/Niko1221](https://github.com/Niko1221/Strata/blob/a0116b2782a5b82622358c04a06496485d12b2d6/setup.py#L228))
```sh
./setup.sh --yes --family unsloth --model UD-IQ4_XS --no-start && ./run-unsloth-ud-iq4_xs.sh
```
3. Windows (cmd): set up this size
```sh
START-HERE.bat --yes --family unsloth --model UD-IQ4_XS --no-start
```
4. Then start it
```sh
run-unsloth-ud-iq4_xs.bat
```
5. And open http://127.0.0.1:8080 (an OpenAI-compatible API at /v1)
Other versions: IQ3_S (12 GB VRAM + 62 GB RAM, 83.6 GB download, NVIDIA or AMD card only) `./setup.sh --yes --family qwen --model IQ3_S --no-start && ./run-iq3_s.sh` · IQ3_XXS (12 GB VRAM + 60 GB RAM, 75.8 GB download, NVIDIA or AMD card only) `./setup.sh --yes --family qwen --model IQ3_XXS --no-start && ./run-iq3_xxs.sh` · IQ2_XS (12 GB VRAM + 48 GB RAM, 68 GB download, NVIDIA or AMD card only) `./setup.sh --yes --family qwen --model IQ2_XS --no-start && ./run-iq2_xs.sh` · Q2_0 (12 GB VRAM + 48 GB RAM, 66.4 GB download, NVIDIA or AMD card only) `./setup.sh --yes --family qwen --model Q2_0 --no-start && ./run-q2_0.sh`
Model repository and documentation: [huggingface.co/Qwen](https://huggingface.co/Qwen/Qwen3.8-Flash-Next)
## Published speed measurements
Reference results for the hardware, engine and settings below. These are not verified speeds of the downloads above or predictions for your machine.
hardware precision speed engine and settings
AMD Ryzen AI Max+ 395 (Stri… UD-IQ4_XS 27.16 tok/s llama.cpp c841aee (build_number 1…
NVIDIA GeForce RTX 5070 12 … Q2_0 · experts … 87.33 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … Q2_0 · experts … 92.97 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … Q2_0 · experts … 81.75 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … Q2_0 · experts … 76.23 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … Q2_0 · experts … 73.73 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … IQ2_XS · expert… 79.62 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … IQ2_XS · expert… 78.6 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … IQ2_XS · expert… 76.34 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … IQ2_XS · expert… 63.69 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … IQ2_XS · expert… 62.72 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … IQ3_XXS · exper… 61.94 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … IQ3_XXS · exper… 61.64 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … IQ3_XXS · exper… 58.49 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … IQ3_XXS · exper… 57.21 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … IQ3_XXS · exper… 48.95 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … IQ3_S · experts… 52.4 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … IQ3_S · experts… 53.3 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … IQ3_S · experts… 48.31 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … IQ3_S · experts… 46.25 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 5070 12 … IQ3_S · experts… 45.51 tok/s Strata 0.1.26, CUDA · Strata 0.1.…
NVIDIA GeForce RTX 3060 12 … Q2_0 · experts … 42.4 tok/s Strata 0.1.41 (official Windows r…
NVIDIA GeForce RTX 3060 12 … Q2_0 · experts … 42.9 tok/s Strata 0.1.41 (official Windows r…
NVIDIA GeForce RTX 3060 12 … Q2_0 · experts … 41.8 tok/s Strata 0.1.41 (official Windows r…
NVIDIA GeForce RTX 3060 12 … Q2_0 · experts … 38.6 tok/s Strata 0.1.41 (official Windows r…
NVIDIA GeForce RTX 3060 12 … IQ2_XS · expert… 42.4 tok/s Strata 0.1.41 (official Windows r…
NVIDIA GeForce RTX 3060 12 … IQ2_XS · expert… 45.5 tok/s Strata 0.1.41 (official Windows r…
NVIDIA GeForce RTX 3060 12 … IQ2_XS · expert… 40.8 tok/s Strata 0.1.41 (official Windows r…
NVIDIA GeForce RTX 3060 12 … IQ2_XS · expert… 37.6 tok/s Strata 0.1.41 (official Windows r…
NVIDIA GeForce RTX 3060 12 … IQ3_XXS · exper… 25.6 tok/s Strata 0.1.41 (official Windows r…
NVIDIA GeForce RTX 3060 12 … IQ3_XXS · exper… 26.2 tok/s Strata 0.1.41 (official Windows r…
NVIDIA GeForce RTX 3060 12 … IQ3_XXS · exper… 27.1 tok/s Strata 0.1.41 (official Windows r…
NVIDIA GeForce RTX 3060 12 … IQ3_XXS · exper… 27 tok/s Strata 0.1.41 (official Windows r…
NVIDIA GeForce RTX 3060 12 … IQ3_S · experts… 18.9 tok/s Strata 0.1.41 (official Windows r…
NVIDIA GeForce RTX 3060 12 … IQ3_S · experts… 20.2 tok/s Strata 0.1.41 (official Windows r…
NVIDIA GeForce RTX 3060 12 … IQ3_S · experts… 20.2 tok/s Strata 0.1.41 (official Windows r…
NVIDIA GeForce RTX 3060 12 … IQ3_S · experts… 18.9 tok/s Strata 0.1.41 (official Windows r…
Each speed links to its source, which has the full settings.
[interactive version](https://whatmodel.sh/model/qwen3-8-flash-next) · [JSON](https://whatmodel.sh/api?format=json&all=1)