other
3B parameters
Commercial OK
Reviewed May 2026

Orpheus 3B 0.1 FT

LLaMA-architecture 3B model fine-tuned as a TTS that emits SNAC audio tokens. Designed for highly expressive, emotion-controllable speech with laughter, sighs, and other paralinguistic markers via inline tags. Apache-2.0 with multiple voice presets.

License: apache-2.0·Context: 0 tokens
BLK · VERDICT

Our verdict

OP · Fredoline Eruo|VERIFIED MAY 29, 2026
unrated

The most interesting open TTS architecture of 2025 — LLM-as-TTS unlocks emotion control that Kokoro/Piper can't match. Worth the 3B footprint if expressivity is the requirement.

Overview

LLaMA-architecture 3B model fine-tuned as a TTS that emits SNAC audio tokens. Designed for highly expressive, emotion-controllable speech with laughter, sighs, and other paralinguistic markers via inline tags. Apache-2.0 with multiple voice presets.

Strengths

  • Best-in-class expressivity — laughter, sighs, emotion tags supported natively
  • Apache-2.0 license, fully commercial
  • Reuses standard LLM inference stack (vLLM, llama.cpp branches) for serving
  • Streamable token-by-token like a text LLM

Weaknesses

  • 3B params — GPU effectively required; CPU is too slow for realtime
  • English-only base checkpoint
  • Requires SNAC audio decoder as a second model in the pipeline
  • Newer project — tooling and benchmarks less mature than Whisper/Kokoro

Quantization variants

Each quantization trades model quality for file size and VRAM. Q4_K_M is the most popular starting point.

QuantizationFile sizeVRAM required
Q4_K_M1.7 GB3 GB

Get the model

HuggingFace

Original weights

huggingface.co/canopylabs/orpheus-3b-0.1-ft

Source repository — direct quantization required.

Hardware that runs this

Cards with enough VRAM for at least one quantization of Orpheus 3B 0.1 FT.

Compare alternatives

Models worth comparing

Same parameter band, plus what's one tier above and below — so you can decide what actually fits your hardware.

Frequently asked

What's the minimum VRAM to run Orpheus 3B 0.1 FT?

3GB of VRAM is enough to run Orpheus 3B 0.1 FT at the Q4_K_M quantization (file size 1.7 GB). Higher-quality quantizations need more.

Can I use Orpheus 3B 0.1 FT commercially?

Yes — Orpheus 3B 0.1 FT ships under the apache-2.0, which permits commercial use. Always read the license text before deployment.

What's the context length of Orpheus 3B 0.1 FT?

Orpheus 3B 0.1 FT supports a context window of 0 tokens (about 0K).

Source: huggingface.co/canopylabs/orpheus-3b-0.1-ft

Reviewed by RunLocalAI Editorial. See our editorial policy for how we research and verify model claims.

Related — keep moving

Before you buy

Verify Orpheus 3B 0.1 FT runs on your specific hardware before committing money.