hjLabs AI Playground
Starting the on-device runtime…
SmolLM2 135M Instruct is a 135M-parameter ONNX model that generates text a token at a time in response to a prompt, streaming its reply as it decodes. On this page it runs entirely inside your browser: the 186 MB download comes straight from the Hugging Face repository “HuggingFaceTB/SmolLM2-135M-Instruct” into your browser cache, and every inference after that executes on your own GPU through WebGPU. That is a one-off download of well under a minute on a normal connection, and small enough that integrated graphics handle it without complaint. It is an English-language model; expect quality to fall off sharply on other languages.
The quickest way to see a real LLM run on your GPU — loads in seconds and replies instantly, but keep the questions simple. In practice people reach for a model like this for drafting, rewriting, summarising a pasted passage, or simply having a conversation that no server ever sees.
Available precisions: q4 (4-bit), q8 (8-bit) and fp32 (full precision). q4 comes first because ONNX Runtime executes 4-bit block-quantised matmuls natively on WebGPU, which is what makes decoding fast enough to be usable — it is a speed choice at least as much as a size one. It is this playground's default pick for the task, and Hugging Face records about 1.4M downloads a month for the repository — a good sign that it is maintained and that the ONNX export is not abandoned.
Because there is no inference server behind this page, there is nothing to rate-limit and nothing to bill: the only cost of running SmolLM2 135M Instruct here is the one-off 186 MB download, and once that is cached the page keeps working with the network switched off.
smollm2 135m instructsmollm2 135m instruct onnxsmollm2 135m instruct browserrun smollm2 135m instruct locallyhuggingfacetb/smollm2-135m-instructchat modelsmollm2 135m instruct webgpusmollm2 135m instruct download sizefree chat modeltransformers.js model
About 186 MB for the q4 build, including the tokenizer and config files the runtime also has to fetch — this is the whole payload, not just the weights file. It is downloaded once and then served from the browser Cache Storage API.
Yes, through the WebAssembly fallback, which is slower but works anywhere. WebGPU (Chrome or Edge 113+, Safari 26+) is what makes it feel interactive. On WASM, prefer the smaller quantized build.
No. HuggingFaceTB/SmolLM2-135M-Instruct is a public repository, so your browser fetches the ONNX files directly over HTTPS. There is no account, no key and no server in between.
No. The model weights are downloaded from Hugging Face to your browser once, and every inference after that runs on your own GPU through WebGPU. Your text, images and audio are never sent anywhere — hjLabs.in has no inference server and no way to see your input.