AI21 Labs has launched the Jamba 1.5 family of open models: Jamba 1.5 Mini and Jamba 1.5 Large, redefining long-context AI with unmatched speed, quality, and multilingual capabilities for global enterprises.

AI21 Labs has taken a major step into the ai ​​landscape by launching the Jamba 1.5 Open Model Familywhich includes Jamb 1.5 Mini and Jamb 1.5 LargeThese models, based on the novel SSM-Transformer architecture, represent a major advancement in ai technology, particularly in handling long-context tasks. AI21 Labs aims to democratize access to these powerful models by publishing them under the Jamba Open Model License, encouraging widespread experimentation and innovation.

Main features of the Jamba 1.5 models

One of the most notable features of the Jamba 1.5 models is their ability to handle exceptionally long contexts. They feature an effective context window of 256,000 tokens, the longest on the market for open models. This feature is critical for enterprise applications that require analysis and summarization of large documents. The models also excel in retrieval-augmented generation (RAG) and agent workflows, improving both the quality and efficiency of these processes.

In terms of speed, Jamba 1.5 models are up to 2.5 times faster over long contexts than their competitors and maintain superior performance across all context lengths within their size class. This speed advantage is crucial for businesses that need fast response times for tasks such as customer support or large-scale data processing.

The quality of the Jamba 1.5 models is another area where they outshine their peers. The Jamba 1.5 Mini has been recognized as the strongest open-top model in its size class, achieving a score of 46.1 in the Arena Hard benchmark, beating larger models such as the Mixtral 8x22B and Command-R+. The Jamba 1.5 Large goes even further, scoring 65.4, which beats leading models such as the Llama 3.1 70B and 405B. This high-quality performance across multiple benchmarks underscores the Jamba 1.5 models’ strength in delivering reliable and accurate results.

Multilingual support and developer readiness

In addition to their technical capabilities, Jamba 1.5 models are designed with multilingual support, covering languages ​​such as Spanish, French, Portuguese, Italian, Dutch, German, Arabic and Hebrew. This makes them versatile tools for global companies operating in diverse language environments.

For developers, Jamba 1.5 models offer native support for structured JSON output, function calls, document object digestion, and citation generation. These features make models adaptable to a variety of development needs, allowing for seamless integration into existing workflows.

Deployment and efficiency

AI21 Labs has ensured that the Jamba 1.5 models are accessible and deployable across multiple platforms. They are available for immediate download on Hugging Face and are supported by major cloud service providers including Google Cloud Vertex ai, Microsoft Azure, and NVIDIA NIM. The models are expected to be available soon on additional platforms such as amazon Bedrock, Databricks Marketplace, Snowflake Cortex, and others, making them easily deployable across multiple environments including on-premises and virtual private clouds.

Another key advantage of Jamba 1.5 models is their resource efficiency. Based on a hybrid architecture that combines the advantages of Transformer and Mamba architectures, these models offer lower memory footprints, allowing companies to manage large contexts on a single GPU. AI21 Labs’ novel quantization technique, ExpertsInt8, further improves this efficiency, optimizing model performance without compromising quality.

Conclusion

AI21 Labs’ launch of the Jamba 1.5 family marks a significant advancement in handling large contexts. These models set new benchmarks in speed, quality, and efficiency and democratize access to cutting-edge ai technology through their open model license. As enterprises continue to look for ai solutions that deliver real-world value, the Jamba 1.5 models stand out as powerful tools capable of meeting the demands of complex, large-scale applications. Their availability on multiple platforms and support for multilingual environments further enhance their appeal, making them a versatile choice for developers and enterprises.

Take a look at the Jamb 1.5 mini, Jamb 1.5 largeand Details. All credit for this research goes to the researchers of this project. Also, don't forget to follow us on twitter.com/Marktechpost”>twitter and join our Telegram Channel and LinkedIn GrAbove!. If you like our work, you will love our fact sheet..

Don't forget to join our Over 49,000 ML subscribers on Reddit

Find upcoming ai webinars here