Flintr is in development.

flintr
Back the launch

Model-neutral

Every frontier model. One subscription.

No house model pushed on you, and no exclusive deal deciding what you can pick. Every model is billed at its provider's list price.

Claude

Anthropic

Anthropic's Claude models are the most popular choice for agentic coding: long multi-file edits, careful refactors and following project rules to the letter.

  • Claude Opus
  • Claude Sonnet
  • Claude Haiku

OpenAI

OpenAI

OpenAI's GPT-5 family is a strong all-rounder for debugging, algorithmic work and explaining unfamiliar code, with a fast mini model for everything quick.

  • GPT-5
  • GPT-5 mini

Gemini

Google

Google's Gemini models pair very large context windows with fast responses, which makes them a good fit for questions that span an entire repository.

  • Gemini Pro
  • Gemini Flash

DeepSeek

DeepSeek

DeepSeek's open-weight models offer strong coding and reasoning performance at a fraction of frontier prices, which stretches your included usage a long way.

  • DeepSeek Chat
  • DeepSeek Reasoner

Mistral

Mistral AI

Mistral AI builds models specifically for code, including Codestral for fill-in-the-middle completion and Devstral for agentic software engineering.

  • Codestral
  • Devstral
  • Mistral Large

Qwen

Alibaba

Alibaba's Qwen Coder models are among the strongest open-weight coding models available, with good agentic skills and long context.

  • Qwen Coder
  • Qwen

Llama

Meta

Meta's Llama models are the most widely used open-weight family. Use the hosted versions in your plan, or connect an endpoint your team already runs.

  • Llama

Kimi

Moonshot AI

Moonshot AI's Kimi models combine long context with strong tool use, which suits agents that need to read a lot before they write.

  • Kimi

Grok

xAI

xAI's Grok models are available like every other provider: same list-price billing, same privacy defaults, no special placement. Flintr doesn't rank any provider above the rest.

  • Grok

Your choice

Let Auto decide, or decide yourself.

Auto

Picks a model per request. Quick edits go to fast models, hard problems go to frontier reasoning.

Pin

Choose a model per project or per conversation, and switch mid-conversation whenever you like.

Bring your own key

Add your own API key per provider. Those requests never count against your plan.

FAQ

Model questions.

How quickly do new models arrive?

We aim for within a day of public release. Pro+ and Ultra subscribers will get early access when a provider offers it.

Do all models cost the same?

No. Each request is priced at its provider's list price, so a fast model costs a fraction of a frontier model. Your included usage is measured in dollars, which means it stretches much further on fast models.

What does Auto actually do?

Auto looks at each request, including how much context it needs and how hard the task looks, and picks a model that should handle it well for the lowest cost. Any model you pin always takes priority.

Can I restrict which models my team uses?

Yes. On Teams and Enterprise, admins can set a model allowlist for the whole organisation, and set spend caps per user.

Your editor. Your models. Your bill.

Flintr is in development and funded by the people who'll use it. Back the launch from €5, or from €25 to get into the private beta.

Back the launch
See planned pricing

Coming to macOS 12+ · Windows 10+ · Linux (deb, rpm, AppImage)