Current Flagship Activefrontierllm Proprietary

Jamba 1.6 Large

by AI21 Labs· Released February 2025· Cutoff December 2024

Jamba 1.6 Large is AI21 Labs' flagship model, combining a hybrid SSM-Transformer architecture for high performance and efficiency. It excels at long-context tasks with a 256K token context window and supports structured outputs like JSON mode and function calling. This model is designed for enterprise-grade reasoning, coding, and multilingual applications.

Official Site API Docs

Input cost

$5.00 per 1M tokens

Output cost

$15.00 per 1M tokens

Context window

256K tokens

Max output

4096 tokens

Modalities

text

License

proprietary

Capabilities

Function CallingCode GenerationStreamingJSON ModeMultilingual SupportLong Context UnderstandingReasoning

Best For

Enterprise applications requiring long-context understanding, structured outputs, and efficient inference.

Strengths

Hybrid SSM-Transformer architecture for efficient long-context processing
Strong performance on reasoning and coding benchmarks
256K token context window enabling analysis of large documents

Limitations

No native vision or multimodal support
Smaller parameter count compared to some frontier models
Limited ecosystem and community compared to open-source models

Use Cases

Document analysis and summarization of long reports

Code generation and debugging in software development

Multilingual customer support chatbots

Complex reasoning tasks in legal or financial domains

Structured data extraction from unstructured text

Conversational AI with long conversation history

Enterprise knowledge base querying

Improvements Over Previous Model

Introduced hybrid SSM-Transformer architecture for better efficiency
Context window increased from 200K to 256K tokens
Improved reasoning and coding benchmark scores
Lower latency and faster inference compared to Jamba 1.5
Enhanced multilingual support with broader language coverage

Back to all models