Free AI for Open Source
A grant of 15,000,000 free AI tokens plus Pro-tier API access for maintainers of actively maintained open source projects.
Modelos de IA 100% de código aberto
Todos os modelos auto- hospedados no Free.ai son de código aberto baixo licenzas permisivas (Apache 2.0, MIT, BSD). Non hai licenzas Llama/ Meta. Cremos que a IA debe ser transparente, auditable e libre de usar.
Modelos de linguaxe Chat
| Modelo | Caso de uso | Licenza | Fonte |
|---|---|---|---|
| Qwen 2. 5- 72B | Chat xeral, razoamento, análise | Apache 2.0 | HuggingFace |
| Qwen 2. 5- 7B | Chat rápido, tarefas lixeiras | Apache 2.0 | HuggingFace |
| Mistral-7B | Chat rápido, de uso xeral | Apache 2.0 | HuggingFace |
| Phi- 3 | Razoamento eficiente, matemáticas | MIT | HuggingFace |
Modelos de código
| Modelo | Caso de uso | Licenza | Fonte |
|---|---|---|---|
| Qwen 2. 5- Coder- 32B | Xeración de código, depuración | Apache 2.0 | HuggingFace |
| DeepSeek-Coder-V2-Lite | Completación do código, revisión | MIT | HuggingFace |
Modelos de imaxe e vídeo
| Modelo | Caso de uso | Licenza | Fonte |
|---|---|---|---|
| FLUX.1-fast | Xeración rápida de imaxes | Apache 2.0 | HuggingFace |
| Stable Diffusion XL | Xeración de imaxes de alta calidade | OpenRAIL++ | HuggingFace |
| Kandinsky 2.2 | Xeración artística de imaxes | Apache 2.0 | HuggingFace |
| CogVideoX-2B | Xeración de vídeo a partir de texto | Apache 2.0 | HuggingFace |
Speech & Audio Models
| Modelo | Caso de uso | Licenza | Fonte |
|---|---|---|---|
| Kokoro | Texto- a- voz | Apache 2.0 | HuggingFace |
| Piper | Texto- a- voz | MIT | GitHub |
| MeloTTS | Texto- a- voz | MIT | GitHub |
| Chatterbox | Texto- a- voz | MIT | HuggingFace |
| OuteTTS | Texto- a- voz | Apache 2.0 | HuggingFace |
| faster-whisper | Fala- a- texto | MIT | GitHub |
| Cera | Fala- a- texto | Apache 2.0 | GitHub |
| ACE-Step 3.5B | Xeración de música | Apache 2.0 | HuggingFace |
| Mel-Band RoFormer | Separación musical (vocais, instrumentais) | MIT | arXiv |
| SCNet | Separación musical (batería, baixo, outros) | Apache 2.0 | arXiv |
Tradución, visión e outros modelos
| Modelo | Caso de uso | Licenza | Fonte |
|---|---|---|---|
| MadLAD- 400 3B | Tradución (máis de 450 linguas) | Apache 2.0 | HuggingFace |
| Real-ESRGAN | Aumento da imaxe | BSD | GitHub |
| BRIA RMBG 2. 0 | Eliminar o fondo | Apache 2.0 | HuggingFace |
| EEUU | Segmentación da imaxe | Apache 2.0 | GitHub |
| DeOldify | Coloración da foto | MIT | GitHub |
| all- MiniLM- L6- v2 | Texto incorporado | Apache 2.0 | HuggingFace |
| Teseracto | OCR | Apache 2.0 | GitHub |
Programa de mantemento de código aberto
Maintainers of actively maintained open source projects can apply for a free grant of 15,000,000 tokens, the token allowance of a year of Pro, plus API keys at the Pro request tier (100,000 requests per month). The grant covers every tool that runs on our self-hosted models. Premium external models are not part of the grant: they use purchased tokens.
Quen califica
- Mantedes o proxecto (propietario ou committer regular), e o seu repositorio é público baixo unha licenza de código aberto aprobada por OSI
- Mantéñense activamente, con confirmacións ou publicacións nos últimos 90 días
- Outras persoas usan o programa, demostrado polo menos por un dos seguintes: 50 estrelas en GitHub, 1000 descargas de paquetes ao mes, ou proxectos que dependen del
- Unha concesión por conta Free.ai nos programas de código aberto, sen ánimo de lucro e de inicio, engadida a unha conta que xa creou
Para solicitar, envíe a ligazón ao repositorio e unha ligazón á proba de uso (estrelas, estatísticas de descarga ou dependentes) desde o enderezo de correo electrónico da súa conta de Free.ai. Aplicar a través do formulario de contacto
FAQ
Yes. Every model we run on our GPU servers is open-source under permissive licenses -- Apache 2.0, MIT, or BSD. We deliberately exclude models with restrictive licenses like Meta's Llama license. You can inspect, use, and modify all models freely.
Meta's Llama license includes restrictions on commercial use for large companies and requires specific attribution. We only use models with truly permissive licenses (Apache 2.0, MIT, BSD) that allow unrestricted commercial and personal use by anyone.
Every model in our table links to its source on HuggingFace or GitHub. You can download the weights directly from these repositories and run them on your own hardware.
Qwen2.5-72B is our most capable chat model for complex reasoning and analysis. Qwen2.5-7B and Mistral-7B are faster options for everyday tasks. Phi-3 is excellent for efficient reasoning and math.
We use five text-to-speech models: Kokoro (Apache 2.0), Piper (MIT), MeloTTS (MIT), Chatterbox (MIT), and OuteTTS (Apache 2.0). Each has different strengths in voice quality, speed, and language support.
FLUX.1-schnell for fast generation (Apache 2.0), Stable Diffusion XL for high-quality images (OpenRAIL++), and Kandinsky 2.2 for artistic styles (Apache 2.0). For video, we use CogVideoX-2B, Wan 2.2 and HunyuanVideo.
If you have developed or know of an open-source model under a permissive license that you think we should host, contact us. We evaluate models based on quality, license compatibility, and user demand.
A free grant of 15,000,000 tokens plus API keys at the Pro request tier (100,000 requests per month) for maintainers of actively maintained open source projects with evidence of use. It covers the tools that run on our self-hosted models. Premium external models are not part of the grant and use purchased tokens. Apply with a link to your repository and to the evidence of use.
We evaluate every model before deploying it based on benchmark performance, real-world output quality, inference speed, and license compliance. Only models that meet our quality bar are added to the platform.
We use MadLAD-400 3B from Google, which supports 450+ languages under an Apache 2.0 license. It handles even low-resource languages that many commercial translation services do not support.
Real-ESRGAN (BSD) for image upscaling, BRIA RMBG 2.0 (Apache 2.0) for background removal, SAM (Apache 2.0) for image segmentation, and DeOldify (MIT) for photo colorization.
Non: os modelos externos premium como GPT-5, Claude e Gemini son propietarios, teñen os seus propios termos de licenza, executan en tokens comprados e non forman parte de ningunha concesión de compensación. Só os modelos autoaloxados listados nesta páxina son de código aberto.