Alibaba
Qwen3.8 Omni Flash
qwen3.8-omni-flash
Total Context
1M
Max Output
131.1K
Released
N/A
Input$0.1143 / M tokens
Output$0.3857 / M tokens
Qwen3.6 35B-A3B
qwen3.6-35b-a3bQwen3.6 35B A3B is a compact active-parameter MoE model: sources describe 35B total parameters with around 3B activated parameters, plus native long context that can be extended toward 1M tokens. Official materials compare it favorably with larger dense models on coding tasks despite the small active footprint. Its value proposition is efficient deployment with surprisingly strong coding and long-context ability.
Context Window
262.1K tokens
Maximum Output
65.5K tokens
Release Date
Apr 17, 2026
Modalities
Qwen3.6 35B-A3B Pricing
| Input Price | Output Price |
|---|---|
| $0.2571/M | $1.5429/M |
Qwen3.6 35B-A3B API Capabilities
Reasoning
Supported
Tool calling
Supported
Temperature parameter
Supported
Attachments
Supported
Knowledge Base
—
Endpoint Protocols
Completions APIMessages APIgemini
Qwen3.6-35B-A3B Model Highlights
Qwen3.6-35B-A3B combines an open-weight sparse MoE architecture with native vision-language processing and improved agentic coding.
Sparse MoE Efficiency
Its hybrid linear-attention and sparse mixture-of-experts design activates about 3B of 35B parameters to improve inference efficiency.
Native Multimodality
A native vision-language design accepts text, images and video and produces text grounded in visual content.
Open-Weight Agentic Coding
The post-trained open-weight model improves frontend workflows and repository-level reasoning while allowing independent deployment.
Qwen3.6-35B-A3B Use Cases
Qwen3.6-35B-A3B supports independently deployed coding agents, repository analysis and visual reasoning over images or video.
Self-Hosted Coding Agent
Deploy the open weights in a controlled environment to generate, revise and debug code through iterative agent workflows.
Repository Code Analysis
Trace relationships across repository files, explain implementation structure and prepare coordinated modification plans.
Visual Media Analysis
Analyze images or video to locate objects, interpret spatial relationships and return a structured textual account.
How to Use Qwen3.6 35B-A3B via the TokenHub API
import OpenAI from "openai"
const client = new OpenAI({
apiKey: process.env.TOKENHUB_API_KEY,
baseURL: "https://us-api.tokenhub.com/v1",
})
const result = await client.chat.completions.create({})
console.log(result.choices[0]?.message?.content)Qwen3.6 35B-A3B Benchmarks
| Index score | ||
|---|---|---|
| Artificial Analysis Intelligence Index | Artificial Analysis broad capability aggregate | 33 |
| Artificial Analysis Coding Index | Artificial Analysis software task aggregate | 35.2 |
| Knowledge & Reasoning | ||
| GPQA | Advanced science problem solving | 84.1% |
| HLE | Broad expert-level exam set | 20.2% |
| Coding & Engineering | ||
| SciCode | Scientific coding challenges | 35.8% |
| Terminal-Bench Hard | Hard terminal task execution | 34.8% |
| Instruction Following & Agent Tasks | ||
| IFBench | Prompt constraint adherence | 64.4% |
| AA-LCR | Long-context reasoning | 63.7% |
| τ²-Bench | Agent workflow tasks | 95.3% |
Metrics sourced from Artificial Analysis
Qwen3.6 35B-A3B Guides & Articles
Qwen 3.6 35B A3B FAQ
Qwen 3.6 35B A3B: capabilities, use cases, limits, and TokenHub guidance.
What does Qwen 3.6 35B A3B focus on?+
Qwen 3.6 35B A3B is a Alibaba Qwen model for open-model multimodal reasoning and efficient deployment.
Which projects fit Qwen 3.6 35B A3B?+
Best for self-hosted deployment, image and video understanding and routine coding assistance, especially when deployment control is the priority.
What is special about Qwen 3.6 35B A3B?+
Key strength: an open MoE variant with a small active-parameter footprint and hybrid thinking that can switch between deliberate and direct responses.
When is another model better?+
Open deployment requires infrastructure, serving, and evaluation work. For stable production behavior matters, consider Qwen 3.6 Plus.
How do I avoid ID mistakes?+
Use TokenHub's exact ID; hosted behavior may differ from self-hosting.
Ready to use Qwen3.6 35B-A3B?
Use one API key to access Qwen3.6 35B-A3B and more AI models through TokenHub.
Media and Discussions
Selected public videos and posts related to this model.
X (Twitter)
Qwen ·
⚡ Meet Qwen3.6-35B-A3B：Now Open-Source！🚀🚀 A sparse MoE model, 35B total params, 3B active. Apache 2.0 license. 🔥 Agentic coding on par with models 10x its active size 📷 Strong multimodal perception and reasoning ability 🧠 Multimodal thinking + non-thinking modes https://t.co/UMiChPaLid
https://pbs.twimg.com/media/HGB1pOdbIAA_kyz.jpgView originalQwen ·
LM Performance：Qwen3.6-35B-A3B outperforms the dense 27B-param Qwen3.5-27B on several key coding benchmarks and dramatically surpasses its direct predecessor Qwen3.5-35B-A3B, especially on agentic coding and reasoning tasks. https://t.co/PyXDNruoy2
https://pbs.twimg.com/media/HGB1yASbcAEzWkd.jpgView originalUnsloth AI ·
2-bit Qwen3.6-35B-A3B did a complete repo bug hunt with evidence, repro, fixes, tests and a PR writeup. 🔥 Run it locally in Unsloth Studio with just 13GB RAM. 2-bit Qwen3.6 GGUF made 30+ tool calls, searched 20 sites and executed Python code. GitHub: https://github.com/unslothai/unsloth https://t.co/1YwMxrWOYr
https://pbs.twimg.com/amplify_video_thumb/2044849663774425088/img/fNsqbHmhlS3uC_47.jpgView originalReddit
Qwen3.6-35b-a3b seems like the best coding model rn
JSVD2 ·
what are your thoughts on this and real life cases and example?
View originalQwen 3.6 35b a3b is INSANE even for VRAM-constrained systems
Lucerys1Velaryon ·
Read full post
Apologies for the 1345th post glazing QWEN but it has literally been a game-changer for me. I’m relatively new to running local LLMs, but the appeal of having a private AI assistant for coding and experimentation is significant. As a software engineer, I’ve also embraced the recent “vibe-coding” trend, mostly using AI to build hobby Android apps in my spare time. Over the past month, I created a web scraper app designed to extract images from ad-heavy websites that are otherwise frustrating to navigate. The project was entirely AI-generated. I didn’t write the code myself, though I did design the architecture and suggest optimizations, which AI still struggles with. I’ll admit I haven’t reviewed most of the code; it’s messy, but since it’s purely a personal hobby project, that’s acceptable. Until recently, I relied on services like Antigravity and Codex, which worked well enough on free plans. However, tighter usage quotas pushed me toward local models. My hardware is modest by local LLM standards: an AMD 7700 XT, 32GB DDR4 RAM, and a Ryzen 5 5600. I experimented with Gemma 3, Gemma 4, and Qwen 2.5 Coder (mostly Q3/Q4 quantizations under 20GB due to VRAM constraints), using LM Studio as the backend with various frontends like GitHub Copilot, RooCode, Cline, Android Studio AI Chat, and OpenCode. Unfortunately, none of these models fit my workflow well. They struggled with even minor bug fixes, frequently exhausted context windows, got stuck in reasoning loops, or failed tool calls repeatedly. Then I tried Qwen 3.6 35B-A3B. Initially, I expected little, but I installed the i1-q4_k_s quant, offloaded all 40 layers to GPU, configured 128k context, enabled flash attention, and used Q8_0 KV quantization. For testing, I gave it a practical task: fix the scraper logic for a problematic website. Gemini Flash (in Antigravity) and MinMax (the free version in Opencode) had both failed to solve this issue despite multiple attempts. Using LM Studio as an OpenAI-compatible endpoint with GitHub Copilot in VS Code, I let it run. It took about 25 minutes, but it succeeded in one shot. With a single initial prompt, it analyzed the site’s HTML structure, compared it against my Kotlin scraper code, and resolved three critical bugs without a single failed tool call. That result was impressive enough that I gave it a second challenge: update the project README with real app screenshots by driving an Android emulator, selecting screenshots based on specific criteria. Using RooCode this time, the process took around 45 minutes. I had to teach it some emulator workflow conventions, such as taking screenshots after every action and analyzing them to track app state, but once instructed, it executed flawlessly. For the first time, I feel like I have a local model capable of reliably handling most of my coding tasks, while reserving cloud-based premium models for more demanding work. Qwen has genuinely made local AI coding practical for me.
Qwen3.6 35B-A3B on a Laptop: My Zero to One Moment
rolznz ·
Read full post
Hi everyone, I'm new here - because I only have a laptop and I only just realized local models are actually good enough now. So I'd like to share my experience, in case it helps others, and also to learn from the more experienced people here. This is the first model that works for me on my ASUS Zenbook Pro 14 (RTX 4060 8GB VRAM, 64GB RAM): fast enough: ~27TPS generation speed at 32k context, or ~18TPS at 256k context smart enough: it can read and write files, use skills, execute CLI commands, use git, follow instructions, and act as a useful thinking partner. Why it's important to me For me this is important because it's where I unconsciously decided to draw the line - that I didn't want to share private information or more personal thoughts with cloud models (even TEE ones). I know I can still get hacked and my data leaked, but for me that's different than giving it up from the first prompt. So for the first time, I now have this fully local, second brain. For me, it's a game changer. I still use cloud models for public stuff I'm still using cloud models for public projects, but for brainstorming and simple personal projects, local is now good enough for me. I'm also now looking into a more powerful desktop machine where maybe I can do some more serious coding. I have had a taste and I want more 😄 Now whenever I see Claude's black box "✽ Envisioning… (41s · ↓ 2.9k tokens · thinking some more with high effort)" it's so frustrating. I have no idea if it's going in the right direction. (whether this is an "efficient" way to do things is another story) My issues so far with Qwen3.6 Qwen3.6 35B A3B is not perfect, here are some minor issues I observed, which I can work around: It makes some mistakes, but normally recovers on its own. Very occasionally it does get stuck in a loop. It does need some human monitoring, which is fine for me. It sometimes doesn't read a skill in full or make the best decision even when it can fit it in context. It seems to sometimes be "lazy". It is very non-deterministic. I didn't do any tweaks here though (because normally it ends up with the result I need). I guess some of these could be improved if I used a larger quantization. My setup For inference I use llama.cpp, with unsloth's Qwen3.6-35B-A3B-UD-IQ3_XXS.gguf. For my harness, I use Pi with pi-llama-cpp extension. The harness runs in multipass and connects to the host running llama.cpp. I've also connected it to my phone through an E2EE Matrix chat (a custom one I built off of pi-messenger-bridge) - although it means I have to keep my laptop on all the time, which is annoying. Another reason for buying another machine which I'm more comfortable to run 24/7. llama.cpp flags for 256k context(18tps): ./build/bin/llama-server -m Qwen3.6-35B-A3B-UD-IQ3_XXS.gguf -ngl 24 -np 1 -fa on -ctk q4_0 -ctv q4_0 -c 262144 --host 0.0.0.0 --port 8088 -ncmoe 32 --no-mmap --jinja llama.cpp flags for the 32k context (27tps): ./build/bin/llama-server -m Qwen3.6-35B-A3B-UD-IQ3_XXS.gguf -ngl 99 -np 1 -fa on -ctk q4_0 -ctv q4_0 -c 32000 --host 0.0.0.0 --port 8088 -ncmoe 32 --no-mmap --jinja What was your Zero to One moment?
YouTube
Qwen3.6 35B-A3B Full Test – Is THIS the Best LOCAL Model Yet?
Bijan Bowen
View originalQwen3.6-35B-A3B + OpenClaw - Agentic Coding Locally for Free
Fahd Mirza
View originalQwen3.6 Local Test | Can it Beat Gemma 4? | Coding, OCR, Image Understanding with llama.cpp | 🔴 Live
Venelin Valkov
View original