LLMSpeed.dev: feel how fast a local LLM runs before buying the hardware
A Show HN tool that simulates tokens-per-second and time-to-first-token for consumer GPUs, Apple Silicon Macs and other rigs, so buyers can preview local LLM speed before spending on hardware.
Why it matters
Attacks a real, expensive pain point — buying a GPU or Mac that turns out too slow for local models — though the site itself admits many figures are projected from specs or similar hardware rather than directly measured.
What you could build with it
A hardware retailer or PC builder could embed the estimator on laptop and GPU product pages so shoppers see which models will run smoothly on a machine before buying, cutting returns driven by mismatched performance expectations.
Does it hold up?
Too early to judge — launched on Show HN today; its speeds are simulated estimates, and no systematic comparison against directly measured results on the same hardware has been published yet.
Built with LLMSpeed.dev
- Talk Around Town: LLMSpeed.dev Claims to Let You Test-Drive Local AI Speed Before Buying HardwareThe Tech Holler · Skeptical deep-dive (published Oct 17) verifying the Show HN post; warns simulated tok/s figures are directional guidance only, and flags AI-generated dev.to posts that invent features the site does not have.
- LLM Updates (October 2026) — AI Model Releases & Provider Activitylmmarketcap · Release tracker listing 'Show HN: Llmspeed – feel how fast a local LLM runs before you buy the hardware' among Oct 17 AI news.
- Llmspeed Benchmark: Pick the Perfect On-Device LLM Hardware Fastdevto · Hands-on walkthrough of the llmspeed CLI: install, run benchmarks per device backend, batch configs, and TCO/electricity cost calculations.
Learn more
First spotted on article: source.
More AI models
OpenAI publishes 722 AI-generated math manuscripts from an unreleased modelOpenAI released 722 mathematical manuscripts in 372 result families to GitHub under Apache-2.0, produced by an…models · JEV 0.74StepFun Step-5-Preview — open BF16 weights releasedStepFun published the full BF16 weights for its flagship Step-5-Preview agentic model (600B-parameter sparse MoE, 27B…models · JEV 0.73ElevenLabs Eleven v4 + v4 Turbo: new emotive TTS architecture with a 100ms real-time variantElevenLabs released Eleven v4 and Eleven v4 Turbo on September 28: a new text-to-speech architecture with inline…models · JEV 0.72Mistral Large 4 (Le Chonk): 1.05T-param open-weight MoEMistral's new flagship: a 1.05T-parameter mixture-of-experts multimodal model with 49B active params and 1M-token…models · JEV 0.72VoiceStudio: open-source, fully-local ElevenLabs alternative with voice cloning, dubbing and dictation in 646 languagesVoiceStudio is an open-source, fully-local voice AI studio: voice cloning from a 3-second sample, voice design, video…models · JEV 0.71NVIDIA Nemotron-pretraining-Indic-v1: 2.8T-token open Indic pretraining corpusNVIDIA with IIT Madras's AI4Bharat released a ~2.8 trillion-token pretraining corpus covering 22 Indian languages plus…models · JEV 0.69
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.