Recursive

What is Recursive?

Recursive is an AI research company, incorporated as Recursive Superintelligence, Inc., with offices in San Francisco and London. Its stated aim is software that improves its own AI research, and it uses an internal system called the Eureka Machine to propose, implement and test research ideas. Its first public model is Flint, a mixture-of-experts model for agentic coding on local hardware, with 40 billion total parameters and 2.5 billion active per token. The weights are on Hugging Face under the Apache 2.0 license, so you can run Flint on your own machine, keep data local and avoid API fees. It works with llama.cpp, Ollama, Unsloth, LM Studio, vLLM and SGLang, and Recursive also released rec, a small inference engine tuned for Flint on recent Apple chips. Recursive also operates hosted APIs and a developer console, but no price list is published.

Use cases

• Run a coding model locally
• Use an agentic coding assistant offline
• Fine-tune an open-weights model
• Serve a model with vLLM or SGLang
• Keep source code on your own machine
• Study automated AI research methods

Who it's for

• Developers
• AI researchers
• ML engineers
• Privacy-conscious teams

Why use Recursive

• Open weights let you run and adapt the model yourself
• Local use keeps your code and data on your own device
• No per-token API fees when you run Flint on your hardware
• Works with several popular local inference tools

Recursive features

• Flint: mixture-of-experts coding model with 40B total and 2.5B active parameters
• Open weights under Apache 2.0 on Hugging Face
• Runs on Apple Silicon and NVIDIA RTX Spark hardware
• Works with llama.cpp, Ollama, Unsloth, LM Studio, vLLM and SGLang
• rec inference engine for M4 Pro and M5 Pro/Max chips
• NVFP4 4-bit quantization
• Published documentation with setup examples

Recursive integrations

• llama.cpp
• Ollama
• Unsloth
• LM Studio
• vLLM
• SGLang

How to get started with Recursive

1. Download the Flint weights from Hugging Face.
2. Load them in a supported runtime such as Ollama, LM Studio, llama.cpp, vLLM or SGLang, or use the rec engine on a recent Mac.
3. Use it as a local coding model, giving explicit instructions and splitting larger work into subtasks.
4. Adapt the model for your own applications under the Apache 2.0 license.

Recursive limitations

• The terms say inputs and outputs sent to the hosted API may be used to train Recursive's models
• The vendor advises breaking complex tasks into smaller subtasks for Flint
• Hosted offerings may be free, pre-release or paid and can change or end at any time

Recursive pricing

Recursive is open source, so you can use and self-host it for free.

Prices can change and may vary by country or exclude tax. Check the official pricing page.

Recursive alternatives

The main Recursive alternatives listed on Debuly are Ollama, LM Studio, Rapid-MLX, TextGen (Text Generation Web UI), Anyscale and FastChat. The table compares pricing and platforms; order is A–Z.

1. 

   Anyscale

   Managed platform for running Ray workloads, with pay-as-you-go compute and options to deploy in your own cloud.

   Developer platformsAutomation and agents
   PaidVisit Anyscale (official site)
2. 

   FastChat

   Open platform for training, serving and evaluating LLM chatbots, with an OpenAI-compatible API and the code behind Chatbot Arena.

   Developer platformsChat assistantsCoding
   Open sourceVisit FastChat (official site)
3. 

   LM Studio

   Desktop app for running open language models on your own computer, now with the Bionic agent, a local API and optional US-hosted cloud models.

   Chat assistantsDeveloper platformsCodingProductivity
   FreemiumVisit LM Studio (official site)
4. 

   Ollama

   Open-source app and CLI for running open AI models locally on macOS, Windows and Linux, with optional paid cloud models.

   Developer platformsCodingChat assistants
   Open sourceVisit Ollama (official site)
5. 

   Rapid-MLX

   Free Apache 2.0 local LLM server and Mac app for Apple Silicon with OpenAI and Anthropic-compatible APIs for coding agents.

   Developer platformsCodingAudio and voice
   Open sourceVisit Rapid-MLX (official site)
6. TT

   TextGen (Text Generation Web UI)

   Open-source desktop and web app for running large language models locally, with chat, vision, tools and an OpenAI-style API.

   Developer platformsChat assistants
   Open sourceVisit TextGen (Text Generation Web UI) (official site)

Recursive FAQ

Sources & verification

Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.

Image credits

• Logo: from framerusercontent.com · fetched 24 Oct 2026 · Vendor trademark, used only to identify the product; removed on request
• Cover: none collected yet, so a generated tile is shown.

Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.