# higgs-audio-v3-tts-4b-transformers by Apolinário from multimodal Source: https://savrn.com/models/higgs-audio-v3-tts-4b-transformers Markdown alternate of the page above; the site index is https://savrn.com/llms.txt --- ## Runs On What it takes to serve higgs-audio-v3-tts-4b-transformers (4.7B parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them. | Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits | | --- | --- | --- | --- | --- | --- | | 16-bit | 9.3 GB | 11.2 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 8-bit | 4.7 GB | 5.6 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 4-bit | 2.3 GB | 2.8 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the [SAVRN Index](https://savrn.com/ai-index/pricing/gpus), read Oct 23, 2026. [higgs-audio-v3-tts-4b-transformers on every accelerator the SAVRN Index prices, at every precision](https://savrn.com/models/higgs-audio-v3-tts-4b-transformers/gpus) ## Model Card A trustremotecode packaging of bosonai/higgs-audio-v3-tts-4b that loads with plain transformers (no SGLang). The weights are the original checkpoint, copied unchanged; only a small modeling.py / configuration.py pair and an automap were added. The model is a standard Qwen3-4B backbone plus a fused multi-codebook audio embedding/head. Reference-audio encoding and waveform decoding use the transformers-native bosonai/higgs-audio-v2-tokenizer (higgsaudiov2tokenizer), loaded automatically on first use. Requires transformers >= 5.5. generatespeech returns a mono 24 kHz waveform as a CPU float32 tensor [L]. - Generation uses Higgs' delay pattern across 8 codebooks (vocab 1026, incl. BOC/EOC… Excerpt from the card by Apolinário from multimodal AI art, licensed other. ## Configuration Architecture HiggsMultimodalQwen3ForConditionalGeneration Context length (tokens) 32,768 Layers 36 Hidden size 2,560 Feed-forward size 9,728 Attention heads 32 Key/value heads 8 Head dimension 128 Vocabulary size 151,936 Model type higgs_multimodal_qwen3 ## Identity and Version Repository multimodalart/higgs-audio-v3-tts-4b-transformers Publisher Apolinário from multimodal AI art Task Text to speech Modality Audio Library transformers Parameters 4.7B parameters Languages en Revision 30f01593ee6a12efa586c92455afe4b76e45095d First published 2026-06-05 Last updated 2026-06-05 ## Files and Weights 11 files, 9.3 GB in total. The weights are 1 file totalling 9.3 GB in safetensors. Weights1 file · 9.3 GB Configuration4 files · 111.2 KB Tokenizer2 files · 11.4 MB Documentation2 files · 28.1 KB Other1 file · 2.4 KB Repository1 file · 1.6 KB Every file | File | Type | Size | SHA-256 | | --- | --- | --- | --- | | model.safetensors | Weights | 9.3 GB | 2f7965264c36 | | config.json | Configuration | 1.7 KB | — | | configuration_higgs_multimodal_qwen3.py | Configuration | 3.3 KB | — | | model.safetensors.index.json | Configuration | 90.1 KB | — | | modeling_higgs_multimodal_qwen3.py | Configuration | 16.1 KB | — | | LICENSE | Documentation | 25.7 KB | — | | README.md | Documentation | 2.4 KB | — | | chat_template.jinja | Other | 2.4 KB | — | | .gitattributes | Repository | 1.6 KB | — | | tokenizer.json | Tokenizer | 11.4 MB | eb883de2de5a | | tokenizer_config.json | Tokenizer | 1.9 KB | — | ## License and Download License other Access Open weights, no gate Download size 9.3 GB [Download from Apolinário from multimodal AI art](https://huggingface.co/multimodalart/higgs-audio-v3-tts-4b-transformers) Released by Apolinário from multimodal AI art through its official repository on Hugging Face. ## Built From - Derived from [bosonai/higgs-tts-3-4b](https://savrn.com/models/higgs-tts-3-4b) ## Memory Requirements | Precision | Weights in memory | | --- | --- | | As published | 9.3 GB | | 16-bit | 9.3 GB | | 8-bit | 4.7 GB | | 4-bit | 2.3 GB | Weights only, from the published parameter count; the key-value cache and runtime add to this. ## Questions About higgs-audio-v3-tts-4b-transformers ### How much GPU memory does higgs-audio-v3-tts-4b-transformers need? About 11.2 GB at 16-bit and 2.8 GB at 4-bit: the weights (4.7B parameters) plus a working margin. A long context needs more. ### What is the cheapest GPU to run higgs-audio-v3-tts-4b-transformers on? At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists. ### What license is higgs-audio-v3-tts-4b-transformers released under? other, as its publisher declares it. Read the license text before commercial use. ### What is higgs-audio-v3-tts-4b-transformers's context length? 32,768 tokens, from the maximum position embeddings in its published configuration. ## Similar Models Model · Text to speech ### [higgs-tts-3-4b](https://savrn.com/models/higgs-tts-3-4b) [Boson AI](https://savrn.com/model-publishers/bosonai) Higgs TTS 3 is built for voice chat: it speaks, not just reads. It turns model responses into expressive conversational speech across 100+ languages, with zero-shot voice cloning and inline control over emotion, style, prosody, pauses, and sound effects. Higgs autoregressive decoder consumes interleaved text and audio tokens. Audio is encoded by the Higgs Tokenizer into 8 codebooks at 25 fps, staggered via a delay pattern, then mapped to backbone hidden states through a multi-codebook fused embedding. Output codes pass through a multi-codebook fused head, are de-delayed, and decoded back to waveform. The model reaches single-digit WER/CER on 102 languages, which split into · Chichewa/Nyanja… Open weights other 4.7B parameters 32,768 tokens transformers [View model](https://savrn.com/models/higgs-tts-3-4b) Model · Text to speech ### [s2-pro](https://savrn.com/models/s2-pro) [Fish Audio](https://savrn.com/model-publishers/fishaudio) Fish Audio S2 Pro is a leading text-to-speech (TTS) model with fine-grained inline control of prosody and emotion. Trained on over 10M+ hours of audio data across 80+ languages, the system combines reinforcement learning alignment with a dual-autoregressive architecture. The release includes model weights, fine-tuning code, and an SGLang-based streaming inference engine. S2 Pro builds on a decoder-only transformer combined with an RVQ-based audio codec (10 codebooks, ~21 Hz frame rate) using a Dual-Autoregressive (Dual-AR) architecture: - Slow AR (4B parameters): Operates along the time axis and predicts the primary semantic codebook. - Fast AR (400M parameters): Generates the remaining 9… Open weights other 4.6B parameters [View model](https://savrn.com/models/s2-pro) Model · Text to speech ### [MOSS-TTS-Local-Transformer-v1.5](https://savrn.com/models/moss-tts-local-transformer-v1-5) [OpenMOSS](https://savrn.com/model-publishers/openmoss-team) MOSS-TTS-Local-Transformer-v1.5 is continued from MOSS-TTS-Local-Transformer-v1.0. It preserves the main 1.0 capabilities, including zero-shot voice cloning, long-form speech generation, token-level duration control, Pinyin/IPA pronunciation control, multilingual synthesis, and code-switching. For the full 1.0 feature walkthrough, input schema, and evaluation tables, please refer to the MOSS-TTS-Local-Transformer-v1.0 README. Compared with MOSS-TTS-Local-Transformer-v1.0, v1.5 focuses on the following improvements: - Stronger multilingual synthesis with language tags: when the language field is omitted, v1.5 may improve some languages and regress slightly on others compared with 1.0. When… Open weights apache-2.0 4.6B parameters 32,768 tokens transformers [View model](https://savrn.com/models/moss-tts-local-transformer-v1-5) Model · Text to speech ### [orpheus-3b-0.1-ft](https://savrn.com/models/orpheus-3b-0-1-ft) [Canopy Labs](https://savrn.com/model-publishers/canopylabs) 03/18/2025 – We are releasing our 3B Orpheus TTS model with additional finetunes. Code is available on GitHub: CanopyAI/Orpheus-TTS Orpheus TTS is a state-of-the-art, Llama-based Speech-LLM designed for high-quality, empathetic text-to-speech generation. This model has been finetuned to deliver human-level speech synthesis, achieving exceptional clarity, expressiveness, and real-time streaming performances. Check out our Colab (link to Colab) or GitHub (link to GitHub) on how to run easy inference on our finetuned models. Do not use our models for impersonation without consent, misinformation or deception (including fake news or fraudulent calls), or any illegal or harmful activity. By… Access requested at publisher apache-2.0 3.8B parameters transformers [View model](https://savrn.com/models/orpheus-3b-0-1-ft) Model · Text to speech ### [higgs-tts-2-3b-base](https://savrn.com/models/higgs-tts-2-3b-base) [Boson AI](https://savrn.com/model-publishers/bosonai) Check our open-source repository https://github.com/boson-ai/higgs-audio for more details! 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. Despite having no post-training or fine-tuning, Higgs TTS 2 excels in expressive audio generation, thanks to its deep language and acoustic understanding. On EmergentTTS-Eval, the model achieves win rates of 75.7% and 55.7% over "gpt-4o-mini-tts" on the "Emotions" and "Questions" categories, respectively. It also obtains state-of-the-art performance on traditional TTS benchmarks like Seed-TTS Eval and Emotional Speech Dataset (ESD). Moreover, the model… Open weights other 5.8B parameters 2,048 tokens transformers [View model](https://savrn.com/models/higgs-tts-2-3b-base) Model · Text to speech ### [step-audio-editx-8bit-mlx](https://savrn.com/models/step-audio-editx-8bit-mlx) [App Automaton](https://savrn.com/model-publishers/appautomaton) This repository contains a self-contained pure-MLX int8 conversion of Step-Audio-EditX for local voice cloning and expressive audio editing on Apple Silicon. All pipeline components are stored as.safetensors — no PyTorch, ONNX, or NumPy files are required at inference time. This bundle is self-contained — all weights are packaged in one repository. Five-stage pipeline, all running pure MLX with bf16 activations: 1. Step1 LM (3.5B params, int8) — autoregressive dual-codebook token generation 2. CampPlus (bf16) — speaker embedding extraction from reference audio 3. Flow conditioner (bf16) — conditions generation on speaker embedding 4. Flow model (int8) — flow-matching mel spectrogram… Open weights apache-2.0 3.5B parameters mlx [View model](https://savrn.com/models/step-audio-editx-8bit-mlx) ## Apolinário from multimodal AI art [All models and datasets](https://savrn.com/model-publishers/multimodalart) ## Versions - [30f01593ee6a](https://savrn.com/models/higgs-audio-v3-tts-4b-transformers/versions/30f01593ee6a) · current 2026-09-18 ## Explore More - [All text to speech models](https://savrn.com/models/tasks/text-to-speech) - [Model comparisons](https://savrn.com/models/comparisons) - [The model directory](https://savrn.com/models) - [Open model prices by host](https://savrn.com/ai-index/pricing/open-models) ## Source - Repository metadata, read 2026-10-20. - [Hugging Face record](https://huggingface.co/multimodalart/higgs-audio-v3-tts-4b-transformers) - [How the hub is built](https://savrn.com/model-hub/methodology)