Model reference · open weights
thinksound is an open-weight audio or speech model from ilintar, listed in the AxForge catalogue. AxForge can bring it up on EU-owned hardware for you on request — with the licence handled where one is required.
About
ThinkSound — GGUF GGUF weights for ThinkSound (text-to-sound-effect), converted for use with thinksound.cpp — a standalone C++/GGML runtime (CUDA + Vulkan), and with Lemonade as a GGML media backend. Source model: https://huggingface.co/FunAudioLLM/ThinkSound License: see the source model. Files DiT and T5 in bf16 (recommended) and q8 (smallest); plus metaclip-text-f32, vae-f32, and the clip/t5 tokenizers.
Summarised from the published model card. Read the full card on the HuggingFace links below.
Specifications
| Maker | ilintar |
|---|---|
| Type | Audio & music |
| Variants | 1 |
| Released | 2026-06-28 |
| Popularity | 3k downloads / month |
| Likes | 1 |
| Licence | Commercial licence needed |
How it works
Variants
Open weights ship in several sizes and precisions. One page, all the variants — pick the one that fits your GPU. VRAM figures are estimates from model size.
| Variant | Params | Precision | VRAM | Fits 16 GB | Weights |
|---|---|---|---|---|---|
| thinksound-gguf | — | GGUF | — | — | Weights ↗ |
Using it via the API
Once AxForge deploys thinksound for you, it answers on the OpenAI-compatible API — the same base URL and keys as every other model. (thinksound below is illustrative; you get the exact model name on deployment.)
$ curl -sS https://api.axforge.ai/v1/audio/transcriptions \ -H "Authorization: Bearer $AXFORGE_API_KEY" \ -F model="thinksound" -F file=@audio.mp3
Details
Tags
Licence
The weights are open but its licence needs a commercial agreement for business use. AxForge can arrange that licence and host the model for you — you pay AxForge, we settle with the model’s maker. Ask us for a quote. Read the licence ↗