Reka AI

Name

Reka AI

Legal name

Reka AI, Inc.

Website

https://reka.ai

Company type

Private

Founded year

2022

Operating status

Operating

Headcount range

11–50 employees

Ownership category

akta.pro rank

Product category

Multimodal AI Platform

NAICS

Software Publishers (51321), Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (5182), Computer Systems Design and Related Services (5415)

SIC

Services-Prepackaged Software (7372), Services-Computer Programming Services (7371), Services-Computer Integrated Systems Design (7373)

akta.pro primary industry

Foundation Model Developers (LLM/Multimodal Model Labs) (HDAAACAA)

akta.pro secondary industries

Open-Source Model Ecosystems & Model Marketplaces (HDAAACAM), Model Hosting, Serving & Inference Platforms (HDAAACAB)

Keywords

• Multimodal AI models
• Vision language models
• AI inference infrastructure
• Enterprise AI platform
• Video understanding AI

Headquarters

HQ city

Sunnyvale

HQ country

United States

HQ region

North America

Offices4 records

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Personnel, Infrastructure, Marketing or Sales, Operations

Revenue model

1. Pay-as-you-go API usage credits: Customers pre-purchase usage credits redeemable against Reka API calls (including Reka Research, Reka Edge, Reka Flash, Reka Vision, Reka Speech) on a consumption basis via Stripe. Unused credits remain in the account and expire after one year.
2. Free / freemium developer tier: Reka Hub, Reka Vision playground, Reka Edge playground, and free API keys (link.reka.ai/free) provide no-cost access that drives product-led adoption; per terms, free-tier content may be used to train models.
3. Enterprise business terms / contracts: Enterprise customers are governed by separate Business Terms (reka.ai/business-terms/), supporting negotiated agreements, on-premise and air-gapped deployments, and large-scale defense/public-sector deals routed via Oracle Defense Ecosystem and Carahsoft.
4. Managed / on-premise deployment services: Reka Vision is sold and deployed as a managed intelligence layer (cloud, on-premise, VPC, or air-gapped) with integration, deployment, and ongoing support services for surveillance, public safety, and industrial customers.
5. Self-hosted / on-device model licensing: Open-weights Reka Edge model distributed via Hugging Face and runnable on customer infrastructure (NVIDIA Jetson, Apple Silicon, Linux/Windows PCs, Qualcomm Snapdragon) with optional Reka Quant for memory-constrained hardware.

Pricing tiers

Go-to-market motion5 records

Core offering

Reka AI develops and sells native multimodal foundation models and the surrounding enterprise infrastructure to deploy them. Its product lineup includes frontier multimodal language models (Reka Core, Reka Flash / Flash 3.1), a 7B open-weight edge vision-language model (Reka Edge), a video understanding and search platform (Reka Vision), a speech transcription/translation model (Reka Speech), a research agent (Reka Research), an enterprise AI workforce platform (Reka Nexus), a video clipping tool (Reka Clip), an inference engine/API (Infer), and a frontier training-data infrastructure (Claru). The portfolio is delivered through consumption-based APIs, free developer playgrounds, enterprise contracts, on-premise/air-gapped deployments, and self-hosted open-weight models.

Differentiator

Problem solved

Functional benefit

Brands

• Reka Edge: Frontier-level edge intelligence VLM optimized for physical AI applications and on-device deployment.
• Reka Vision
• Reka Labs
• Reka Core
• Reka Flash
• Reka Research
• Reka Speech
• Reka Nexus
• Reka Clip
• Reka Quant
• Space

Products and services

• Reka Edge Open-weights 7B vision-language model for on-device multimodal reasoning; targets robotics, automotive, AR wearables, drones, and ADAS developers needing sub-second visual intelligence on memory-constrained hardware.
• Reka Core Reka's frontier-class multimodal language model spanning text, image, video, and audio understanding; consumed via the Reka API and used inside Reka Nexus and partner platforms such as Snowflake.
• Reka Flash / Reka Flash 3.1 Efficient multimodal language model with reasoning upgrades in Flash 3 and reinforcement-learning + Reka Quant improvements in Flash 3.1; available via Reka API, OpenRouter, Hugging Face, and partner channels.
• Reka Vision Multimodal AI video understanding and visual search platform offering persistent video indexing, natural-language search, contextual alerts, highlight and clip generation, and visual reasoning. Sold as a managed intelligence layer (cloud, on-premise, VPC, or air-gapped) to defense, public safety, media, and enterprise security customers.
• Reka Research Research agent that browses the web and private documents to answer complex questions, with Parallel Thinking that generates N candidate responses in parallel and resolves them with a verifier model.
• Reka Speech High-throughput 850M-parameter multilingual speech-to-text and speech-to-speech translation model with timestamps; targeted at enterprise transcription and multilingual workflows.
• Reka Nexus Enterprise AI workforce platform powered by Reka's multimodal reasoning model; orchestrates tools, APIs, and enterprise data sources for intelligence tasks.
• Reka Clip AI-powered video clipping tool that generates short-form content from long-form video via natural-language prompts; targeted at broadcasters and content creators.
• Infer Inference engine and API for multimodal AI, built for speed, scale, and enterprise reliability. Routes calls to Reka Edge, Reka Flash, Reka Core, Reka Research, Reka Vision, and Reka Speech, and is available via API, MCP, and the Reka app.
• Claru Data infrastructure that builds the training data behind frontier AI, including egocentric video, robotics trajectories, world-model footage, and expert human judgment at scale.

Named customers3 records

Segments6 records

Ideal customer profiles6 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

Integration9 records

AI capability7 records

Feature8 records

Scale indicators12 records

Recent moves6 records

Expansion highlights8 records

Direct peers

• AI21 Labs: AI21 Labs develops the Jamba family of foundation models with hybrid SSM-Transformer architecture, sold via API and on-prem to enterprise. Comparable mid-sized foundation model lab competing for enterprise AI platform budgets.
• OpenAI: OpenAI develops frontier multimodal foundation models (GPT-4o, Sora) offered via API to enterprise and developers — same product category, modality stack, and consumption-based pricing model as Reka Core, Flash, Vision, and Speech.
• Mistral AI: Mistral AI develops frontier language and multimodal models, including open-weight releases on Hugging Face and a paid API. Directly comparable open-weights-plus-API strategy to Reka Edge, with similar European enterprise positioning.
• Twelve Labs: Twelve Labs builds multimodal AI specialized for video understanding, offering video embedding, search, and generation APIs. Closest direct competitor to Reka Vision in the long-form video understanding and search category.
• Cohere: Cohere develops enterprise-focused foundation models (Command, Embed) delivered via API and on-premise deployments. Comparable enterprise GTM, on-prem capability, and partner-led distribution to Reka's Business Terms and Snowflake partnership motion.
• Anthropic: Anthropic builds Claude, a multimodal foundation model sold to enterprise via API and cloud partnerships (AWS, Google Cloud). Comparable frontier-model positioning, enterprise GTM, and safety/governance emphasis to Reka's business.

Others

• Hugging Face: Hugging Face operates the leading open-source model marketplace and inference platform. Functions as both a distribution channel for Reka Edge's open weights and an indirect competitor via the broader community of multimodal models it hosts.

Broad incumbents

• Meta AI (Llama): Meta develops the open-weight Llama family of foundation models with multimodal variants. Competes with Reka Edge on the open-weights axis, though Meta's distribution extends to billions of consumer endpoints across its apps.
• Google DeepMind (Gemini): Google DeepMind develops Gemini, a natively multimodal foundation model spanning text, image, video, and audio, distributed via Vertex AI. Overlaps with Reka's multimodal stack but operates at hyperscaler scale across Google's broader consumer and cloud portfolio.
• SambaNova Systems: SambaNova provides full-stack AI infrastructure (custom chips, models, cloud) for enterprise deployment of generative AI. Overlaps with Reka Infer and on-premise enterprise deployment motion at the AI-platform layer.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat7 records

Key risks6 records

Key highlights7 records

Customer concentration

• GitHubgithub.com/reka-ai
• LinkedInlinkedin.com/company/reka-ai
• Xx.com/RekaAILabs

Revenue estimate

Valuation estimate

Number of profiles

Profiles7 records

Subsidiaries3 records

Funding overview

Funding rounds2 records

Investors6 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A1 record

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Reka AI