FlexAI

What is FlexAI?

FlexAI is a French AI infrastructure company that runs open models in production. Its Token Factory provides serverless inference across more than 30 open models for text, vision and media, according to FlexAI, through an OpenAI-compatible API with streaming. Beyond serverless, FlexAI offers Dedicated Endpoints and AI Factory for private, customer-controlled deployments.

As Brijesh Tripathi, CEO of FlexAI, puts it in FlexAI's announcement: "Models are going to keep changing. Teams need the freedom to choose what works for their workload today without creating an infrastructure decision they're locked into tomorrow. That's what we're making possible with Eden AI."

FlexAI at a glance

Why FlexAI: open models built for production

Models tested for agent workloads

FlexAI says it tests the models it serves for the behaviours agent workloads depend on, including real tool calls and execution across multiple turns, rather than only checking that a request with a tools field is accepted. Its own line sums it up: swap models, not your agent.

Global when you want flexibility, EU when you need the boundary

FlexAI runs a global Token Factory and a separate EU Token Factory. According to FlexAI, in an EU workspace inference is served on EU infrastructure, prompts, responses and usage stay within the EU boundary, and a request for a model that is not available in the EU is not moved elsewhere. The EU region runs a smaller catalogue.

On Eden AI, the EU route is available today for DeepSeek V4 Flash 0731, Gemma 4 26B A4B and Gemma 4 31B, called with the @eu suffix, for example flexai/gemma-4-31b-it@eu.

Start with tokens, scale without starting over

Teams can start with per-token serverless inference through Eden AI, then move to FlexAI's Dedicated Endpoints when traffic becomes predictable, or to AI Factory for larger private deployments, without changing the application around the model.

When should you choose FlexAI?

Choose FlexAI if:

You are building agents on open models and want a provider that validates tool calling and multi-turn behaviour on the models it serves.

You want low-cost open models for high-volume work, with prices starting at $0.02 per 1M input tokens on Eden AI.

You may need dedicated capacity or a private deployment later, and want to start with the same models on a per-token API.

Consider another provider if:

You need the full catalogue inside the EU today. On Eden AI, three FlexAI models have an EU route so far; the others are served on FlexAI's global infrastructure for now.

You need a proprietary frontier model. FlexAI serves open models only for now.

FlexAI pros and cons

FlexAI models on Eden AI

This is a selection. The live list of FlexAI models with current pricing is in the Eden AI catalogue: GET https://api.edenai.run/v3/models is public and needs no key.

Try FlexAI through Eden AI

FlexAI sits behind the same OpenAI-compatible endpoint as every other provider on Eden AI. The model is simply part of the request:

import requests

response = requests.post(
    "https://api.edenai.run/v3/chat/completions",
    headers={"Authorization": "Bearer YOUR_EDENAI_API_KEY"},
    json={
        "model": "flexai/DeepSeek-V4.1-Flash",
        "messages": [{"role": "user", "content": "Hello!"}],
    },
)
print(response.json()["choices"][0]["message"]["content"])

To keep inference in the EU, use an EU route such as flexai/gemma-4-31b-it@eu. One key and one bill cover FlexAI and every other provider on Eden AI.

Frequently asked questions about FlexAI on Eden AI