HeadsUpAI

Qwen Partners with Fireworks AI for Global Access to Qwen 3.6 Plus

Qwen partnered with Fireworks AI, an inference platform for fast model serving, to host the lab's closed-weights models. The collaboration centers on qwen3p6-plus, a flagship Mixture-of-Experts (a design that activates only specific sub-networks per token) model previously restricted to Alibaba's internal cloud infrastructure.
Model
Qwen 3.6 Plus
Pricing (Uncached Input)
$0.50 per 1M tokens
Pricing (Cached Input)
$0.10 per 1M tokens
Pricing (Output)
$3.00 per 1M tokens
Architecture
Mixture-of-Experts
Supported Features
Image Input, Function Calling

This partnership bridges an accessibility gap for Western developers who rely on Fireworks AI's high-volume inference but want to use Qwen's top-tier reasoning. While the lab is famous for Qwen's open-weight releases, the "Plus" series represents its most capable proprietary intelligence, now available globally with enterprise-grade reliability.

You can access the model via a serverless API that supports multimodal (processing text and images together) image input and function calling. Pricing is set at $0.50 per million tokens for uncached input and $3.00 for output, positioning it as a high-performance alternative for Qwen's production-ready agentic coding and complex reasoning workflows.

View the full update on app.fireworks.ai
Qwen
Qwen
@Alibaba_Qwen
X

šŸ“¢ Official Announcement: Qwen Partners with Fireworks AI to Accelerate Access to Qwen Family Models We are pleased to announce a strategic partnership between Qwen and Fireworks AI to deliver optimized, production-ready deployment of Qwen's closed weights models via the Fireworks Platform. @FireworksAI_HQ This collaboration empowers developers and enterprises to: āœ… Deploy Qwen models with lower latency and reduced fine tuning and inference costs āœ… Leverage enterprise-grade reliability, security, and scalability āœ… Integrate seamlessly into modern AI workflows šŸ”¹ Get started with Qwen on Fireworks: https://t.co/SEGxfJAGM4 #Qwen #FireworksAI #OpenSourceAI #LLM #AIInfrastructure #ResponsibleAI #DeveloperCommunity

51retweets801likes
View on X

Still wondering? A few quick answers below.

Qwen 3.6 Plus is Alibaba's latest flagship closed-weights model. It uses a Mixture-of-Experts architecture, which improves efficiency by only activating a subset of its parameters for each task. The model is designed for high-performance reasoning and is now available for deployment outside of Alibaba's own cloud infrastructure through a partnership with Fireworks AI.

No, Qwen 3.6 Plus is a closed-weights model, unlike many other releases in the Qwen family that are open-source. While Alibaba frequently releases open weights for its smaller models, the Plus and Max versions are proprietary. Access to this specific model is provided exclusively through Fireworks AI's platform for users outside of Alibaba's internal ecosystem.

Fireworks AI offers Qwen 3.6 Plus through a serverless API with a pay-per-token pricing model. Uncached input costs $0.50 per million tokens, while cached input is significantly cheaper at $0.10 per million tokens. Output tokens are priced at $3.00 per million. This structure allows developers to scale usage without managing dedicated infrastructure or paying upfront costs.

Qwen 3.6 Plus supports multimodal capabilities, specifically allowing for image inputs alongside text prompts. It also includes native support for function calling, enabling the model to interact with external tools and APIs. Developers can integrate the model using the Fireworks Python client, a REST API, or the OpenAI-compatible Python client for seamless workflow integration.

You can access Qwen 3.6 Plus through the Fireworks AI platform using their serverless API. While the model is available for immediate use on a pay-per-token basis, enterprises requiring higher reliability or specific performance guarantees can contact Fireworks AI to set up dedicated instances. The platform provides documentation and API keys to help developers get started with integration.

Share this update