EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
Hacker News·Aug 12·5 min read
Qwen/Qwen3.8-2.4T-A95B
TL;DR
Qwen3.8-2.4T-A95B introduces a powerful open-model for AI applications, featuring enhanced capabilities like vision input and a 1M context length. It is designed for easy integration with popular frameworks and offers a managed API service for scalable inference.
✦ Why It Matters
Explore the Qwen3.8-2.4T-A95B model today to enhance your AI applications with advanced features.
Key Takeaways
- Utilize Qwen3.8 for complex coding tasks requiring multi-step reasoning.
- Adjust reasoning depth with reasoning_effort for optimal performance.
- Leverage extensive context lengths for detailed outputs in your applications.
- Integrate with frameworks like vLLM and SGLang for seamless deployment.
Full Summary
Qwen3.8-2.4T-A95B is the latest model in the Qwen family, featuring 2.4 trillion parameters and 95 billion activated for enhanced performance in coding and professional tasks. Built on the architecture of Qwen3.5, it introduces significant improvements in autonomous planning and multi-step task execution.
The model supports a context length of up to 1,010,000 tokens, allowing for complex reasoning and detailed outputs. It is compatible with popular frameworks like vLLM and SGLang, facilitating easy integration into existing workflows.
Users can adjust reasoning depth with parameters like reasoning_effort, optimizing for speed or thoroughness. The model's performance is validated through various benchmarks, showcasing its capabilities across diverse tasks.