Model garden Router · warm Dedicated · available

Qwen3.8 Flash Next

Instantly via the EU router or as a dedicated GPU deployment. Data stays in Europe.

[!Note] This repository contains model weights and configuration files for the post-trained model in the Hugging Face Transformers format. These artifacts are compatible with Hugging Face Transformers, vLLM, SGLang, TokenSpeed, etc.

Qwen/Qwen3.8-Flash-Next
text+image->text · Qwen · sovereign EU
Runs on EU infrastructure operated by European companies; US marketplace capacity is never part of this chain. Full chain
180B
Parameters
262K
Context window
432GB
Minimum VRAM
POST /api/v1/chat/completions 200 OK

Specifications

Parameters 180B
Context window 262,144 tokens
Minimum VRAM 432 GB
Architecture Qwen4ExpForConditionalGeneration (vLLM)
License other
Modality text+image->text
Released August 2026
Publisher Qwen ↗

Pricing

Shared router · per token
€0.23
Input (per 1M tokens)
€0.58
Output (per 1M tokens)
€0.06 per million cached input tokens
Dedicated GPU · per hour
from €21,48 per hour
Your own vLLM instance on European cloud (432 GB VRAM), billed hourly.

Shared EU router, pay-per-token, scale-to-zero. Dedicated GPU deployments are billed hourly, see pricing.

Bench index

Bench is our quality index for every model, 0 to 100. It rests on the Epoch AI Capabilities Index (about 40 benchmarks), completed with LiveBench, and is refreshed every morning.

77.2 / 100 · Via LiveBench

Translated from its LiveBench result onto the same scale.

Range 70.9 to 83.5

Sources: Epoch AI ↗ · LiveBench ↗ · Refreshed daily, last on 25-09-2026

LiveBench · 76.2 · Rank 28 of 63
Reasoning 87.4
Coding 72.6
Agentic coding 61.6
Mathematics 85.8
Data analysis 74.2
Language 74.6
Instruction following 77.1
LiveBench · All tasks
Reasoning
theory of mind 80.8
zebra puzzle 98.8
spatial 100.0
logic with navigation 70.0
Coding
code generation 69.0
code completion 76.1
Agentic coding
javascript 68.2
typescript 46.7
python 70.0
Mathematics
AMPS Hard 99.0
integrals with game 59.0
math comp 94.1
olympiad 91.2
Data analysis
consecutive events 75.6
tablejoin 51.0
tablereformat 96.1
Language
connections 90.2
plot unscrambling 55.8
typos 78.0
Instruction following
paraphrase 79.5
simplify 70.3
story generation 81.4
summarize 77.4

Call it now

Drop-in replacement for OpenAI: change only the base URL and API key. The Anthropic format (/v1/messages) is supported too.

curl https://hostyourai.com/api/v1/chat/completions \
  -H "Authorization: Bearer hyai-..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "Qwen/Qwen3.8-Flash-Next",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Frequently asked questions

Can I run Qwen3.8 Flash Next in the EU?

Yes. HostYourAI runs Qwen3.8 Flash Next on GPUs in European datacenters via vLLM. Prompts and outputs never leave the EU and there is no US cloud provider in the chain.

Is hosting Qwen3.8 Flash Next GDPR-compliant?

Yes. All processing happens inside the EU, a Data Processing Agreement (DPA) is available and the subprocessor list is public. Open-source weights also mean: no training on your data.

How much does Qwen3.8 Flash Next cost?

Via the shared EU router you pay €0.23 per million input tokens and €0.58 per million output tokens, with no fixed costs. For high volume or isolation you can also run Qwen3.8 Flash Next as a dedicated hourly GPU instance.

Is the API OpenAI-compatible?

Yes. You use the standard OpenAI SDKs with a custom base URL (https://hostyourai.com/api/v1). The Anthropic Messages API is supported as a drop-in as well.

More models from Qwen

Qwen Image 2.1 PE T2I

We are excited to open-source Qwen-Image-2.1, a unified text-to-image generation and image editing model in the Qwen family. With just 7B parameters in its visual generation component (32 Single-Stream DiT layers), Qwen-Image-2.1 balances generation quality, inference efficiency, and versatility.

9.4B 262K context View model →
Qwen Image 2.1

We are excited to open-source Qwen-Image-2.1, a unified text-to-image generation and image editing model in the Qwen family. With just 7B parameters in its visual generation component (32 Single-Stream DiT layers), Qwen-Image-2.1 balances generation quality, inference efficiency, and versatility.

7.1B View model →
Qwen3.8 27B

[!Note] This repository contains model weights and configuration files for the post-trained model in the Hugging Face Transformers format. These artifacts are compatible with Hugging Face Transformers, vLLM, SGLang, TokenSpeed, etc.

28B 262K context View model →
Qwen3 ASR 0.6B hf

The Qwen3-ASR family includes Qwen3-ASR-1.7B and Qwen3-ASR-0.6B, which support language identification and ASR for 52 languages and dialects. Both leverage large-scale speech training data and the strong audio understanding capability of their foundation model, Qwen3-Omni. The 1.7B version achieves state-of-the-art performance among open-source ASR models and is competitive with the strongest proprietary commercial APIs.

0.8B 66K context View model →
Qwen3 ASR 1.7B hf

The Qwen3-ASR family includes Qwen3-ASR-1.7B and Qwen3-ASR-0.6B, which support language identification and ASR for 52 languages and dialects. Both leverage large-scale speech training data and the strong audio understanding capability of their foundation model, Qwen3-Omni. The 1.7B version achieves state-of-the-art performance among open-source ASR models and is competitive with the strongest proprietary commercial APIs.

2B 66K context View model →
Qwen3.6 27B

[!Note] This repository contains model weights and configuration files for the post-trained model in the Hugging Face Transformers format. These artifacts are compatible with Hugging Face Transformers, vLLM, SGLang, KTransformers, etc.

28B 262K context View model →

Try Qwen3.8 Flash Next for free

Creating an account takes a minute. Test Qwen3.8 Flash Next straight away in the playground.

Start for free