# SmolLM2 135M browser LLM configuration and benchmarks

Canonical page: https://browserllm.dev/models/smollm2-135m/

Provider: Hugging Face. Nominal parameters: 0.135B. [Upstream model card](https://huggingface.co/HuggingFaceTB/SmolLM2-135M-Instruct).

Measurements describe an exact model artifact, quantization, runtime, browser, device, workload and cache state. Performance medians exclude warmup. Tokenizers and actual output lengths differ across models. These tests do not measure intelligence or task quality; no overall model score is assigned. Missing telemetry is unavailable, never zero. CPU is dedicated Chrome process-tree capacity; Chrome GPU time is experimental driver-accounted time including rendering, not GPU utilization. Power is whole-host subsystem power, not model-only power. Browser memory is summed process footprint, not deduplicated RAM.

## Pinned runnable artifacts

- SmolLM2-135M-Instruct-q0f16-MLC: q0f16; WebLLM 0.2.82; context limit 4096 tokens; required shader features shader-f16; revision 06fc409f2212adff29d93cbc742c1d6957302df5. [Pinned model files](https://huggingface.co/mlc-ai/SmolLM2-135M-Instruct-q0f16-MLC/resolve/06fc409f2212adff29d93cbc742c1d6957302df5/).
- SmolLM2-135M-Instruct-q0f32-MLC: q0f32; WebLLM 0.2.82; context limit 4096 tokens; required shader features none beyond WebGPU; revision ba1ecf858a4dfc6ac3e668c388a7088bdfb3c091. [Pinned model files](https://huggingface.co/mlc-ai/SmolLM2-135M-Instruct-q0f32-MLC/resolve/ba1ecf858a4dfc6ac3e668c388a7088bdfb3c091/).

## Published evidence

### q0f16 — 2dff46ad-a429-4f7b-9e10-145915624392

Captured: 2026-10-18T05:48:26.300Z. WebLLM 0.2.82; Chrome 155.0.0.0; Dedicated Chrome · full session · back-to-back · native power + GPU. Protocol: 3 × 40s · 0s pacing. Completed measured requests: 1. Output tokens per request: 63.

- Median runtime-reported decode speed: 61.38 tok/s.
- Median first-output latency: 22.40 ms.
- Median actual response time: 1.06 s.
- Sampled peak Chrome footprint: 3.85 GB.
- [Raw published JSON report](https://browserllm.dev/reports/2dff46ad-a429-4f7b-9e10-145915624392.json).