FlexAI

FlexAI is an AI infrastructure platform that runs managed inference on open models through a single OpenAI compatible API key, with a path from pay as you go serverless use to dedicated endpoints and a private cloud. The company describes it as infrastructure for agent native AI, and its site lets visitors try image generation and chat completion in a playground before creating an account. The platform has three parts. Token Factory is the serverless service, offering more than 30 open models across text generation, code, vision, embeddings, image generation, speech to text and text to speech, such as Gemma, Qwen, Llama, GPT OSS, GLM, DeepSeek and Mistral models. Because it follows the OpenAI API, it works as a drop in replacement for the OpenAI SDK, with streaming, tool calling and reasoning models supported. Managed Services provides reserved dedicated GPUs behind the same API, billed per minute with autoscaling. AI Factory runs the same serving and training stack inside a customer's own boundary, from a virtual private cloud to air gapped environments. AgentOS is a layer for building governed agents, with a UI, CLI, API, SDK, portable skills, scoped memory and model routing. Serverless pricing is per model and published per million tokens, with separate rates for input, cached input and output. For example, Llama 3.1 8B Instruct is $0.02 input and $0.05 output per million tokens, and GPT OSS 120B is $0.03 input and $0.17 output. Overall rates start from $0.01 per million tokens. Dedicated and private cloud capacity is priced separately.