Gated, non-commercial (CC BY-NC 4.0): ask for access on this page. Cilla Mini is the 18M model for CPUs, phones and browsers: Cilla shrunk with Kyutai's small-model recipe (8 layers, a 64-dim interface, a one-step drifting sampler head, the 2.6M Mimi decoder), guidance baked in.
Alternatives to cilla-mini
120 open-weight models that do the same job as cilla-mini (28M TTS), ranked by how close they are in task and size, then by popularity and how recent they are. Licence and VRAM are shown for each.
03/18/2025 – We are releasing our 3B Orpheus TTS model with additional finetunes. Code is available on GitHub: CanopyAI/Orpheus-TTS
3.8B128K ctx Apache 2.049K 765Mar 2025- Same category
- Larger (3.8B vs 28M)
- Commercial use OK
Dia is a 1.6B parameter text to speech model created by Nari Labs. It was pushed to the Hub using the PytorchModelHubMixin integration.
1.6BApache 2.024.4K 135Jun 2025- Same category
- Larger (1.6B vs 28M)
- Commercial use OK
rumik-oss 1 is a 3b multilingual text-to-speech model from rumik ai, trained on fewer than 70,000 hours of speech while performing competitively with existing TTS models. it combines code-switched synthesis, description-conditioned delivery, and inline vocalization control, with 24 khz audio output.
3.4B8K ctx CC BY-NC 4.0897 68Sept 2026- Same category
- Larger (3.4B vs 28M)
Fish Speech V1.4 is a leading text-to-speech (TTS) model trained on 700k hours of audio data in multiple languages.
?BCC BY-NC-SA 4.0409 460Sept 2024- Same category
Fun-CosyVoice 3.0 is an advanced text-to-speech (TTS) system based on large language models (LLM), surpassing its predecessor (CosyVoice 2.0) in content consistency, speaker similarity, and prosody naturalness. It is designed for zero-shot multilingual speech synthesis in the wild.
500MApache 2.015.2K 99Jul 2025- Same category
- Larger (500M vs 28M)
- Commercial use OK
Dia2 is a streaming dialogue TTS model created by Nari Labs.
1.9BApache 2.06.7K 189Nov 2025- Same category
- Larger (1.9B vs 28M)
- Commercial use OK
Key Improvements: - 🔊 Higher Quality: 44.1kHz sampling rate preserves more high-frequency details for better voice cloning - ⚡ More Efficient: Reduced token rate (6.25Hz) lowers computational cost while maintaining performance - 🎓 Fine-tuning Support: Train personalized voice models with SFT or…
802MApache 2.06K 364Dec 2025- Same category
- Larger (802M vs 28M)
- Commercial use OK
Sarashina2.2-TTS is a Japanese-centric text-to-speech system built on a large language model, developed by SB Intuitions. It supports Japanese and English, delivering high pronunciation accuracy, naturalness, and stability across diverse speaking styles, with zero-shot voice generation support.
810M8K ctx Unknown licence2.3K 76Apr 2026- Same category
- Larger (810M vs 28M)
We release IndicF5, a near-human polyglot Text-to-Speech (TTS) model trained on 1417 hours of high-quality speech from Rasa, IndicTTS, LIMMITS, and IndicVoices-R.
351MMIT41.1K 479Mar 2025- Same category
- Larger (351M vs 28M)
- Commercial use OK
Maya-1 is an open-weight text-to-speech model from Maya Research that combines natural-language voice descriptions with inline expressive controls. It generates speech as discrete audio tokens, which are decoded into a 24 kHz waveform.
3.3B128K ctx Apache 2.08.1K 949Oct 2025- Same category
- Larger (3.3B vs 28M)
- Commercial use OK
State-of-the-art Voice AI has been locked behind web APIs for too long. The NeuTTS Nano Multilingual Collection is a collection of super-fast, highly realistic, on-device TTS speech language models with instant voice cloning - built to run smoothly on CPUs and edge devices.
229MCustom licence978 81Nov 2025- Same category
- Larger (229M vs 28M)
CSM (Conversational Speech Model) is a speech generation model from Sesame that generates RVQ audio codes from text and audio inputs. The model architecture employs a Llama backbone and a smaller audio decoder that produces Mimi audio codes.
1.6BApache 2.033.4K 2.5KMar 2025- Same category
- Larger (1.6B vs 28M)
- Commercial use OK
This is the repo for the MARS5 English speech model (TTS) from CAMB.AI.
?BAGPL 3.0509 481Jun 2024- Same category
- Commercial use OK
We are open-sourcing Higgs TTS 2, a powerful audio foundation model pretrained on over 10 million hours of audio data and a diverse set of text data.
5.8BCustom licence9.4K 693Jul 2025- Same category
- Larger (5.8B vs 28M)
State-of-the-art Voice AI has been locked behind web APIs for too long. The NeuTTS Nano Multilingual Collection is a collection of super-fast, highly realistic, on-device TTS speech language models with instant voice cloning - built to run smoothly on CPUs and edge devices.
229MCustom licence310 27Feb 2026- Same category
- Larger (229M vs 28M)
VoxCPM is a novel tokenizer-free Text-to-Speech (TTS) system that redefines realism in speech synthesis. By modeling speech in a continuous space, it overcomes the limitations of discrete tokenization and enables two flagship capabilities: context-aware speech generation and true-to-life zero-shot…
500MApache 2.05.2K 830Sept 2025- Same category
- Larger (500M vs 28M)
- Commercial use OK
Ryuichiro Higashinaka, Shinnosuke Takamichi, and Tetsuji Ogawa. "KABURI-TTS: Phoneme-Keyed Activity-conditioned Bi-channel Utterance Rendering for Interaction." In Proceedings of the 2026 Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC), 2026.
?BApache 2.00 17Jul 2026- Same category
- Commercial use OK
MeloTTS is a high-quality multi-lingual text-to-speech library by MyShell.ai. Supported languages include:
?BMIT699 28Feb 2024- Same category
- Commercial use OK
Voxtral TTS is a frontier, open-weights text-to-speech model that’s fast, instantly adaptable, and produces lifelike speech for voice agents. The model is released with BF16 weights and a set of reference voices.
4BCC BY-NC 4.02.7K 959Nov 2025- Same category
- Larger (4B vs 28M)
Dia2 is a streaming dialogue TTS model created by Nari Labs.
1.1BApache 2.02.3K 27Nov 2025- Same category
- Larger (1.1B vs 28M)
- Commercial use OK
Zonos-v0.1 is a leading open-weight text-to-speech model trained on more than 200k hours of varied multilingual speech, delivering expressiveness and quality on par with—or even surpassing—top TTS providers.
1.6BApache 2.019.7K 439Feb 2025- Same category
- Larger (1.6B vs 28M)
- Commercial use OK
Fish Speech V1.2 is a leading text-to-speech (TTS) model trained on 300k hours of English, Chinese, and Japanese audio data.
?BCC BY-NC-SA 4.0190 210Jul 2024- Same category
See also the pre-print research paper, the project page, the GitHub repository, and the evaluation pipeline.
750MCC BY 4.04.9K 19Jul 2025- Same category
- Larger (750M vs 28M)
- Commercial use OK
Chatterbox-Flash is a block-diffusion zero-shot TTS model that extends the Chatterbox-TTS pipeline with a parallel masked decoder while preserving streaming generation.
?BMIT0 34May 2026- Same category
- Commercial use OK