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.
100% Open-Source AI Models
Mofuta o mong le o mong o sebetseng ka bomong ka Free.ai ke mohloli o bulehileng ka tlase ho litokelo tse lumellanang (Apache 2.0, MIT, BSD). Ha ho Llama/Meta litokelo. Re nahana hore AI e lokela ho ba e hlakileng, e ka hloloheloa, le e bolokehile ho e sebelisa.
Chat Litlhaku tsa Senyesemane
| Mofuta | Ho sebelisa | Laesense | Mofuta |
|---|---|---|---|
| Qwen2.5-72B | Lipuisano tsa kajeno, litlhaku, litlhaku | Apache 2.0 | HuggingFace |
| Qwen2.5-7B | Lipuisano tse potlakileng, mesebetsi e bonolo | Apache 2.0 | HuggingFace |
| Mistral-7B | Lipuisano tse potlakileng, tse sebelisoang ka bongata | Apache 2.0 | HuggingFace |
| Phi-3 | Ho nahana ka ho nepahetseng, lipalo | MIT | HuggingFace |
Li-code Models
| Mofuta | Ho sebelisa | Laesense | Mofuta |
|---|---|---|---|
| Qwen2.5-Coder-32B | Ho theha khoutu, ho tlosa liphoso | Apache 2.0 | HuggingFace |
| DeepSeek-Coder-V2-Lite | Ho kopanya | MIT | HuggingFace |
Litšoantšo le livideo
| Mofuta | Ho sebelisa | Laesense | Mofuta |
|---|---|---|---|
| FLUX.1-speed | Ho theha setšoantšo ka potlako | Apache 2.0 | HuggingFace |
| Stable Diffusion XL | Ho theha setšoantšo se nang le boleng bo phahameng | OpenRAIL++ | HuggingFace |
| Kandinsky 2.2 | Ho theha setšoantšo sa setsebi | Apache 2.0 | HuggingFace |
| CogVideoX-2B | Ho theha video ho tloha ho tekeng | Apache 2.0 | HuggingFace |
Litšoantšo tsa puo le tsa molumo
| Mofuta | Ho sebelisa | Laesense | Mofuta |
|---|---|---|---|
| Kokoro | Text-to-speech | Apache 2.0 | HuggingFace |
| Piper | Text-to-speech | MIT | GitHub |
| MeloTTS | Text-to-speech | MIT | GitHub |
| Chatterbox | Text-to-speech | MIT | HuggingFace |
| TTS | Text-to-speech | Apache 2.0 | HuggingFace |
| faster-whisper | Ho bua ho ea ho ngola | MIT | GitHub |
| Li-wax | Ho bua ho ea ho ngola | Apache 2.0 | GitHub |
| ACE-Step 3.5B | Ho theha lipina | Apache 2.0 | HuggingFace |
| Mel-Band RoFormer | Ho arola lipina (li-vocals, li-instrumental) | MIT | arXiv |
| SCNet | Ho arola lipina (li-drums, bass, tse ling) | Apache 2.0 | arXiv |
Ho fetolela, ho bona le li-models tse ling
| Mofuta | Ho sebelisa | Laesense | Mofuta |
|---|---|---|---|
| MadLAD-400 3B | Ho fetolela (450+ lipuo) | Apache 2.0 | HuggingFace |
| Real-ESRGAN | Ho phahamisa setšoantšo | BSD | GitHub |
| BRIA RMBG 2.0 | Ho tlosa sehlooho | Apache 2.0 | HuggingFace |
| US | Ho arolelana setšoantšo | Apache 2.0 | GitHub |
| Ho tlosa ho qala | Ho lekola litšoantšo | MIT | GitHub |
| all-MiniLM-L6-v2 | Li-embeds tsa mongolo | Apache 2.0 | HuggingFace |
| Tesseract | OCR | Apache 2.0 | GitHub |
Program ea ho boloka lifaele
Batsamaisi ba liprojekte tse sebetsang tse sebetsang ka ho hlaka li ka sebelisana le li-tokens tsa 15,000,000, li-token tse fanoeng ka selemo sa Pro, hammoho le li-API tsa API ka kopo ea Pro (lipotso tsa 100,000 ka khoeli). Likopo li kenyelletsa lisebelisoa tsohle tse tsamaeang ka li-models tsa rona tse sebetseng ka bomong. Li-models tsa Premium tsa kantle ha li be le karolo ea likopo: li sebelisa li-token tse rekotiloeng.
Bao e ka bang
- U boloka projeke (motsamaisi kapa motsamaisi oa kamehla),'me sebaka sa eona sa ho boloka se teng ka ntle ho laesense ea OSI e amohelehang ea mohloli o bulehileng
- E bolokiloe ka ho sa feleng, ka li-commits kapa li-releases ka matsatsi a 90 a fetileng
- Ba bang ba e sebelisa, e bonts'itsoeng ka ho feletseng ka: 50 GitHub stars, 1,000 package downloads a month, or projects that depend on it
- Litšenyehelo tse le 'ngoe ka akhaonto ea Free.ai kahare ho li-open source, li-nonprofit le li-startup programs, li kenyelelitsoe akhaonteng eo u e entseng
Ho kenya, romella sehokela sa polokelo le sehokela sa ho bonts'a ts'ebeliso (li-stars, lipalo-palo tsa ho kenya kapa li-dependent) ho tsoa ho aterese ea imeile ea ak'haonte ea hau ea Free.ai. Sebetsa ka ho sebelisa foromo ea ho ikopanya
Lipotso tse ka etsahalang
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.
Litšenyehelo tsa mahala tsa 15,000,000 tokens le li-API ke li-keys ho Pro request tier (100,000 lipotso ka khoeli) bakeng sa ba bolokang liprojekte tse sebetsang tse kopanetsoeng tse nang le litokomane tsa ho sebelisa. E kenyelletsa lisebelisoa tse tsamaeang ka li-models tsa rona tse kopanetsoeng. Li-models tsa Premium tsa kantle ha li be karolo ea litheko le ho sebelisa li-tokens tse rekiloeng. Sebelisa le link ho repository ea hau le ho litokomane tsa ho sebelisa.
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.
Ha ho joalo: li-models tsa kantle tsa premium joalo ka GPT-5, Claude le Gemini li na le litokelo tsa bona, li na le litaelo tsa bona tsa laesense, li sebetsa ka li-token tse rekiloeng'me ha li na karolo ea likopo tsa likopo. Li-models tse arolelanoang feela tse boletsoeng ka leqephe lena li na le mohloli o bulehileng.