Complete pricing guide for Jamba. Compare all plans, analyze costs, and find the perfect tier for your needs.
Not sure if free is enough? See our Free vs Paid comparison →
Still deciding? Read our full verdict on whether Jamba is worth it →
mo
No credit card required; credits expire after 3 months.
mo
Unlimited seats are listed for the pay-as-you-go plan; usage is billed by token volume.
mo
Unlimited seats are listed for the pay-as-you-go plan; usage is billed by token volume.
mo
Volume discounts and premium API rate limits are available through sales.
Pricing sourced from Jamba · Last verified March 2026
Jamba is used for long-context enterprise AI workflows where teams need to process large documents, internal knowledge bases, or complex records with low latency. The website specifically calls out financial records, contracts, and whole-knowledge-base search as examples for its 256K context window. It is also positioned for finance, technology, defense, healthcare, and manufacturing teams that need secure AI systems. Because it is a model family rather than an end-user app, most teams will use it inside custom applications, agentic workflows, or private AI infrastructure.
Yes. The website explicitly lists self-hosted deployment as an option, using the phrase 'Your data, your infra — your rules.' It also mentions secure cloud deployment with trusted technology partners and private-by-design systems for keeping proprietary data locked down. This makes Jamba relevant for organizations that cannot send sensitive data to a standard public API endpoint. Teams should still confirm the exact hosting package, licensing terms, and operational requirements with AI21 before committing.
The website states that Jamba supports a 256K context window. That is a major part of its positioning for enterprise-grade document processing, especially for lengthy records, contracts, and knowledge-base search. A large context window can reduce the need for aggressive document splitting, although teams still need good retrieval, prompt design, and evaluation practices. In production, performance will also depend on the selected Jamba model, deployment environment, and workload size.
The scraped page lists Jamba2 3B, Jamba2 Mini, and Jamba Reasoning 3B as part of the downloadable model family. Jamba2 3B is described as a compact model for reliability, steerability, on-device applications, and agentic workflows. Jamba2 Mini is positioned for efficient, steerable output on core enterprise workflows. Jamba Reasoning 3B is described as a compact reasoning model with record latency and context-window length for enterprise-grade reasoning.
The current directory pricing value is Freemium, and AI21's pricing page lists a free trial with $10 in credits for 3 months and no credit card required. Published pay-as-you-go API rates include Jamba Mini at $0.20 per 1M input tokens and $0.40 per 1M output tokens, and Jamba Large at $2.00 per 1M input tokens and $8.00 per 1M output tokens. AI21 also states that an average token corresponds to about 1 word or 6 English characters. For managed, private, or self-hosted deployments, teams should expect to request custom pricing or a demo from AI21.
AI builders and operators use Jamba to streamline their workflow.
Try Jamba Now →Paris-based frontier AI lab — open-weight and commercial LLMs (Mistral Small/Large, Codestral, Mixtral), Le Chat assistant with Agent Builder, and La Plateforme for fine-tuning and EU-sovereign hosting.
Compare Pricing →Toronto-based enterprise AI platform: Command family LLMs, Embed and Rerank retrieval models, plus the North agent workspace — built for private, secure, fully customizable deployment in the enterprise.
Compare Pricing →Google's most intelligent AI assistant with multimodal capabilities including text, image, video, and music generation, plus conversational AI and deep integration with Google services.
Compare Pricing →