EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
Hacker News·Aug 17·5 min read
Qwen3.8 27B scores 52 on Artificial Analysis
TL;DR
Qwen3.8 27B is a competitive AI model that excels in intelligence and pricing, scoring 52 on the Artificial Analysis Intelligence Index. It supports both text and image inputs with a large context window of 256k tokens.
✦ Why It Matters
Consider integrating Qwen3.8 27B into your projects for enhanced AI capabilities at a lower cost.
Key Takeaways
- Utilize Qwen3.8 for both text and image processing tasks.
- Leverage its high verbosity for detailed output generation.
- Take advantage of its competitive pricing for cost-effective AI solutions.
- Explore self-hosting options with its open weights availability.
Full Summary
Qwen3.8 27B, developed by Alibaba, is a cutting-edge AI model that integrates text and image processing capabilities. It boasts a significant score of 52 on the Artificial Analysis Intelligence Index, which evaluates models based on reasoning, knowledge, mathematics, and coding, placing it well above the median score of 9.
During its evaluation, Qwen3.8 generated 160 million tokens, indicating a high level of verbosity compared to the median of 43 million tokens. The model features a context window of 256,000 tokens, allowing it to handle extensive text and conversation history in a single request.
Priced at $0.00 per million input and output tokens, it is competitively positioned against similar models. With 27 billion parameters, Qwen3.8 is classified as an open weights model, meaning its weights are publicly available for self-hosting.