Jamba (AI21 Labs)
Israeli production model. Interleaves Mamba SSM layers with Transformer attention (1:7 ratio) + MoE blocks every 2 layers. SSM layers maintain fixed-size recurrent hidden state — compressed long-context memory bypassing growing KV cache.
Active · Verified 2026-10-22 · fresh as of 2026-10-22 · 41 cited facts, 3 with a third-party source · 1 recorded change · Suggest a correction
At a glance
- Type
- SSM-Transformer hybrid + 256K KV-bypass context vendor fix
- Tier
- T1
- Latest release
- Jamba2 3B / Jamba2 Mini (52B, 12B active) (2026-01-08) vendor fix
- Pricing
- Free trial ($10 credit, no CC required); pay-as-you-go per token; enterprise custom pricing with VPC/on-prem vendor fix
- Funding
- $608M total Series D · 2025-05 third-party fix
- Freshness
- 2026-10-22 fix
- Layer
- large-models repo fix
- Licensing
- open-core (estimate) fix
- SDK languages
- python, rest, typescript vendor fix
- Compliance
- iso27001, iso27017, iso27018, iso42001, soc2-type2 vendor fix
When to use
Optimised for: long context (256K) at lower cost via SSM-Transformer hybrid
Anti-fit: not for ultra-low-cost commodity inference (premium positioning)
Pros & cons
Pros
Hybrid Transformer/Mamba architecture means memory cost grows sub-quadratically — long-context inference is genuinely cheaper.
Cons
Smaller ecosystem than Llama / Mistral; less third-party tooling support.
Claims & capabilities
256K context on single 8-GPU node. 3× throughput vs Mixtral 8x7B on long contexts. Jamba Reasoning 3B runs 250K-context inference on a laptop.
Technical surface
- API surface
- searched not found
- Backend storage
- searched not found
- Deployment
- Both (AI21 Studio cloud; Azure/AWS Bedrock/GCP Vertex AI; private VPC; on-premises)
- Embedding model
- searched not found
- Multi-tenancy
- searched not found
- MCP
- not documented publicly
- A2A
- not documented publicly
- OpenTelemetry
- not documented publicly
Similar systems
Other foundation models (substrate reference) in the catalog, ranked by inbound references.
- Anthropic Claude (foundation models) T1
Claude family of frontier foundation models — Sonnet 4 (workhorse), Opus 4.5 (frontier capability), Haiku 4.5 (fast/cheap). Frontier model used as substrate by Claude Code, Anthropic Memory tool, Cline, Aider, Continue.dev, Cursor, Goose, and dozens of other catalog entries. Memory in Claude is parametric (in weights) — discrete episodic memory lives outside the model in the Claude Memory tool (see Platform-provider memory section).
- OpenAI GPT family (GPT-5 / GPT-4o / o3 / o4) T1
OpenAI's foundation model family — GPT-5 (2025-08 flagship), GPT-4o (multimodal workhorse), o3/o4 reasoning series. Substrate for ChatGPT, the OpenAI API, Microsoft Copilot, Azure OpenAI, and many downstream agents in this catalog (Devin, Codex CLI, Operator, Cursor mode). Parametric memory in weights; discrete user memory ships in ChatGPT Memory feature (separate row).
- Google Gemini 3 family T1
Google DeepMind's Gemini 3 frontier model family — Gemini 3 Pro (top tier), 3 Flash (fast), 3 Nano (on-device). Substrate for Google Workspace AI, Vertex AI Agent Platform, Gemini Code Assist, Project Mariner, and Project Astra. Multimodal-native; 1M-10M context window. Largest context window of any frontier model. Used by Gemini Enterprise Agent Platform Memory Bank (separate row).
- Mistral Large 2 / Mixtral family T1
French frontier-model lab (Paris). Family includes Mistral Large 2 (123B dense, Jul-2024), Mixtral 8x22B (open-weights MoE), Mistral Small 3 (24B, open-weights Jan-2025), Codestral (code specialist). EU-headquartered alternative to US/China model labs; GDPR-friendly data residency. Distinct from the Mistral Le Chat memory feature (already in catalog) — this row covers the model family itself as substrate.
- xAI Grok 4 T1
Elon Musk's xAI frontier model — Grok 4 launched 2025-07. Trained on Colossus (100k+ GPU supercluster in Memphis). Deeply integrated with X/Twitter (real-time tweet data access). Grok 4 Heavy variant for premium tier. No frontier-tier open-weights — Grok 1 (314B MoE) is the only weight release. The x.ai site now brands itself SpaceXSI (page title, checked 2026-10-29).
- Cohere Command R+ / Command A T1
Toronto-based foundation-model lab (Aidan Gomez, ex-Google Brain 'Attention Is All You Need' co-author). Command family is enterprise / RAG-optimised: Command A (Mar-2025, 111B dense, 256k ctx), Command R+ (104B MoE), Command R7B (smallest tier). Announced merger with Aleph Alpha April-2026. Open weights via CC-BY-NC (non-commercial); Cohere-hosted for commercial.
Sources (40 cited facts, 3 third-party)
Each fact links to the page it came from; the badge says who published it (how to read sources).
- What it is SSM-Transformer hybrid + 256K KV-bypass context vendor
- What it does & what's distinct Israeli production model. Interleaves Mamba SSM layers with Transformer attention (1:7 r… vendor
- Claims / benefits 256K context on single 8-GPU node. 3× throughput vs Mixtral 8x7B on long contexts. Jamba… vendor
- Modalities text vendor
- Created 2020-11 third-party
- Latest release Jamba2 3B / Jamba2 Mini (52B, 12B active) (2026-01-08) vendor
- License Apache-2.0 (Jamba2); earlier Jamba 1.5–1.7 under Jamba Open Model License registry
- GitHub not applicable — open weights on Hugging Face (huggingface.co/ai21labs) registry
- Funding $608M total Series D · 2025-05 third-party
- Customers / scale $58M ARR 201-500 ppl Boston Consulting Group, Bain & Company, Moderna, Amgen vendor
- Pricing Free trial ($10 credit, no CC required); pay-as-you-go per token; enterprise custom pric… vendor
- Compliance SOC 2; ISO 27001/17/18 certified; GDPR; enterprise deployment enables GLBA + HIPAA-ready vendor
- Data handling Zero data visibility option via private VPC or on-prem; Microsoft Azure does not share p… vendor
- Deployment Both (AI21 Studio cloud; Azure/AWS Bedrock/GCP Vertex AI; private VPC; on-premises) vendor
- HQ Israel (Tel Aviv) vendor
- Founders / pedigree Tel Aviv-based; co-founders Yoav Shoham (Stanford AI professor emeritus) and Amnon Shashua third-party
- Memory volume / scale 256K context on single 8-GPU node vendor
- Consistency strong vendor
- Versioning none vendor
- Schema evolution schema locked vendor
- Forgetting policy LRU eviction vendor
- Import / export Hugging Face checkpoints (Apache-2.0) registry
- Integration count 0 vendor
- Orchestration single-agent vendor
- Programmatic control read-only API + on-prem deploy vendor
- Vendor benchmarks vendor: 256K context single 8-GPU node; 3x throughput vs Mixtral 8x7B on long contexts vendor
- Pricing specifics per-token-stored / per-API-call: free trial $10; pay-as-you-go per token; enterprise VPC… vendor
- Validated verticals financial-services, regulated enterprise (Israel + global) vendor
- Time to running <10min (AI21 Studio API) vendor
- Anti-fit not for ultra-low-cost commodity inference (premium positioning) vendor
- Optimised for long context (256K) at lower cost via SSM-Transformer hybrid vendor
- Adjacent infrastructure self-contained model API; or self-host weights vendor
- Pros Hybrid Transformer/Mamba architecture means memory cost grows sub-quadratically — long-c… vendor
- Cons Smaller ecosystem than Llama / Mistral; less third-party tooling support. vendor
- Project & sources ai21.com/jamba vendor
- architecture-deployment-locus hybrid vendor
- positioning-platform-vs-product platform vendor
- taxonomy-three-layer-rollup large-models repo
- compliance-certifications iso27001,iso27017,iso27018,iso42001,soc2-type2 vendor
- sdk-languages python,rest,typescript vendor
What changed
Changes to this record's lifecycle, licence, release, pricing and hosting facts, from the catalog's git history. Subscribe (Atom)
Related systems
Referenced by (6)
- LangChain (framework) integrates with — LangChain docs have an AI21 integration provider page (AI21 Labs' Jurassic models for text generation). (verified 2026-10-29)
- LiteLLM integrates with — LiteLLM docs have an AI21 provider page. (verified 2026-10-29)
- LlamaIndex (framework) integrates with — LlamaIndex has an official LLM integration package llama-index-llms-ai21 in its repo. (verified 2026-10-29)
- OpenRouter integrates with — OpenRouter lists AI21 as an inference provider (providers API slug "ai21"). (verified 2026-10-29)
- Portkey AI integrates with — Portkey docs have an AI21 LLM provider integration page. (verified 2026-10-29)
- TrueFoundry AI Gateway integrates with — TrueFoundry AI Gateway docs have a AI21 model-provider page. (verified 2026-10-29)