MODEL · RESEMBLE AI · TURBO: 350M · MULTILINGUAL: 0.5B LLAMA BACKBONE
Chatterbox (Turbo + Multilingual v3)
SOTA open-source voice cloning — 5-second reference audio, paralinguistic tags `[laugh]` `[sigh]` `[cough]` native in Turbo, <150 ms latency, ethical PerTh watermarking on by default. The June 10, 2026 Multilingual v3 release keeps the same 0.5B Llama backbone and MIT license while extending coverage to 25 total languages (incl. 4 dialects + 6 tuned Language Packs).
License: MIT · Context: n/a · Released: Base: mid-2025 · Turbo: December 15, 2025 · Multilingual v3 (25 languages): June 10, 2026
The decision in five lines
- The call
- Buy — for voice
- Best for
- voice
- Runs on
- 23 hardware picks fit (cheapest: Intel Arc B580 12 GB · $249)
- Watch out
- Pure narration with no cloning need — Kokoro-82M is 10× lighter and Apache 2.0 (Chatterbox is MIT but still heavier).
- Evidence
- Estimated
- Turbo: 350M · Multilingual: 0.5B Llama backbone
- PARAMETERS
- TTS + VOICE CLONE
- TYPE
- —
- CONTEXT
- ~4–6 GB (Turbo)
- VRAM AT Q4
Where we recommend this
Every tier slot in the planner where this model is a top or alternate pick. Pulled live from planner.js — when the planner refreshes, this table stays current.
The call
SOTA open-source voice cloning — 5-second reference audio, paralinguistic tags `[laugh]` `[sigh]` `[cough]` native in Turbo, <150 ms latency, ethical PerTh watermarking on by default. The June 10, 2026 Multilingual v3 release keeps the same 0.5B Llama backbone and MIT license while extending coverage to 25 total languages (incl. 4 dialects + 6 tuned Language Packs).
When not to use: Pure narration with no cloning need — Kokoro-82M is 10× lighter and Apache 2.0 (Chatterbox is MIT but still heavier).
Runner notes
`pip install chatterbox-tts` or `devnen/Chatterbox-TTS-Server` for OpenAI-compatible REST API. CUDA / ROCm / CPU supported.
Hardware that fits
Every hardware pick whose memory fits this model at the quant we recommend. Sorted cheapest-first — the top row is your best-value fit. Click through for the full buyer’s guide.
- Intel Arc B580 12 GBPerfect · 2.3× 12 GB · $249–$299
- NVIDIA RTX 3060 12 GBPerfect · 2.3× 12 GB · $280–$400
- Minisforum UM890 ProPerfect · 4.7× 32 GB DDR5 (shared) · $463–$580 all-in
- RTX 5060 Ti 16 GBPerfect · 3.1× 16 GB · $560–$610
- AMD Radeon RX 9070 XTPerfect · 3.1× 16 GB · $649–$779
- Mac Mini M4 16 GBPerfect · 2.1× 16 GB unified · $799 (new floor) / $499–$599 (eBay/residuals)
- AMD Radeon RX 7900 XTXPerfect · 4.7× 24 GB · $810 used / ~$1,340 new
- NVIDIA RTX 3090 (used, single)Perfect · 4.7× 24 GB · $950–$1,200
- NVIDIA RTX 5070 TiPerfect · 3.1× 16 GB · $980–$1,300
- NVIDIA RTX 5080Perfect · 3.1× 16 GB · $1,250–$1,400
- MacBook Air M5 24 GBPerfect · 3.1× 24 GB unified · $1,499–$1,899
- Mac Mini M4 Pro 24 GBPerfect · 3.1× 24 GB unified · $1,599
- Dual RTX 3090 (used)Perfect · 9.3× 48 GB · $1,800–$2,500 all-in
- NVIDIA RTX 4090Perfect · 4.7× 24 GB · $2,200–$2,800
- Framework Desktop (Ryzen AI Max+ 395)Perfect · 16.7× 128 GB unified · $2,459–$2,851 (128 GB config)
- M5 Pro MacBook Pro 48 GBPerfect · 6.2× 48 GB unified · $2,999–$3,599
- NVIDIA RTX 5090Perfect · 6.2× 32 GB · $3,500–$4,300
- NVIDIA RTX A6000 (48 GB, used)Perfect · 9.3× 48 GB ECC · $3,500–$4,500
- Mac Studio M4 Max 64 GBPerfect · 8.3× 64 GB unified · $3,799
- NVIDIA DGX SparkPerfect · 16.7× 128 GB unified · $4,699
- M5 Max MacBook Pro 64 GBPerfect · 8.3× 64 GB unified · ~$5,199 (est.; June 25 2026 increase)
- Mac Studio M3 Ultra 96 GBPerfect · 12.5× 96 GB unified · $5,299
- Dual RTX 5090Perfect · 12.4× 64 GB (2×32) · $8,500–$10,500
Next step
Find-by-model — see what hardware runs this→