Determinstic LLM inference for lowest price Gemma 4, with Windows XP

In plain words

Deterministic LLM inference provides open-weight model access through a free preview service, ensuring identical outputs down to the byte across runs. Built on Gemma 4 and compatible with OpenAI clients, it offers streaming, image support, structured outputs, and full sampling parameter control via a web playground. The service targets developers seeking reproducible AI inference with transparent token-based pricing planned for post-preview. Features include tool use, logprobs, seeds, and a tokenization endpoint for advanced users.

written from the facts on this page · September 2026

From the sources

Open-weight models, fully deterministic. The same answer every time, byte for byte. Free while we're in preview. Sign up in a dialog, make a key in API Keys. It's free while we're in preview: no card. OpenAI-compatible: point any client at the base URL with your key. The playground streams answers, takes pictures, and exposes every sampling parameter. Questions: the Contact window. Free during the preview. List prices per million tokens for when billing starts. Streaming, tools, structured outputs, images, logprobs, seeds. A token-ids endpoint for callers that tokenize themselves. Base URL: https://api.tokendelivery.ai/v1 This desktop is a landscape thing, like it's 2001. Rotate to…from tokendelivery.ai

• 
  Open-source engine running Gemma 4 26B in 2 GB RAM on any M-series MacJul 2026 · github.com · ▲919

  Hi HN, I built a specialized inference engine for running 4-bit Gemma 4 26B-A4B-IT on any M-series Mac using about 2 GB of RAM. It is called TurboFieldfare and is written in Swift and Metal. I have always adored on-device AI. It feels like magic that you can run a powerful NN on your Mac or iPhone. So I wanted to push the limits a bit and run a model whose weights don’t fit in memory. The model’s 4-bit quantized weights occupy roughly 14 GB, which makes running it with conventional inference tools almost impossible on an 8 GB or even 16 GB Mac once the OS, applications, and KV cache are…

• 
  Google Gemma 4Apr 2026 · blog.google · ▲437

  Google's most intelligent open models to date

• 
  Speeding up LLM inference 2x times (possibly)2024 · asciinema.org · ▲419

  Here's a project I've been working on for the last few months. It's a new (I think) algorithm, that allows to adjust smoothly - and in real time - how many calculations you'd like to do during inference of an LLM model. It seems that it's possible to do just 20-25% of weight multiplications instead of all of them, and still get good inference results. I implemented it to run on M1/M2/M3 GPU. The mmul approximation itself can be pushed to run 2x fast before the quality of output collapses. The inference speed is just a bit faster than Llama.cpp's, because the rest of implementation…

• 
  AI Observability by OpenObserveAug 2026 · openobserve.ai · ▲317

  OpenTelemetry-native observability for agents and LLMs

• 
  Gemma2024 · deepmind.google · ▲259

  Google’s new state-of-the-art open source LLMs

• 
  We made our own inference engine for Apple Silicon2025 · github.com · ▲186

  We wrote our inference engine on Rust, it is faster than llama cpp in all of the use cases. Your feedback is very welcomed. Written from scratch with idea that you can add support of any kernel and platform.

• 
  Spira Maxima▲578

  Video model that turns script into viral social media videos

  AI · 19d ago · spira.ai

• 
  Rill Browser▲540

  The browser where Claude Code and Codex work beside you

  AI · 19d ago · rill.love

• 
  Germany's new sovereign AI model Kolibri▲424

  Kolibri is Aleph Alpha

  AI · 21d ago · tej.as

• 
  Giving Opus 5.5 a simulated paint canvas▲392

  Paintings by AI models: programs run through a simulation of oil paint.

  AI · 22d ago · stillwet.art

• 
  Beam.win▲379

  Run AI agents on one Mac and use its windows on another

  AI · 5d ago · beam.win

• 
  FastRouter.ai▲364

  Route requests to the right LLM for cost, latency & quality

  AI · 23d ago · fastrouter.ai

• 
  An e-ink frame that hears birds and draws them as 1800s illustrations▲2,389

  E-ink bird frame for Raspberry Pi - real-time bird detection by audio, fully local AI, rendered as real, hand-cut 1800s bird illustrations. - arnegiacomo/fugleramme

  AI · Sep 2026 · github.com

• WI
  What if the speed of light was 5 km/h?▲629

  Special relativity simulator in which light crawls along at a human speed, so Lorentz contraction, aberration and Doppler shift happen at walking pace.

  Life & fun · Sep 2026 · rivendell.dmitrybrant.com

• 
  Jev▲544

  Fast, structured AI decisions for software automation

  AI · Sep 2026 · console.typesafe.ai

• 
  ReSO AI▲510

  Get discovered & recommended when customers ask AI

  AI · Sep 2026 · resollm.ai

• 
  Kilo Code for iOS and Android▲498

  Start coding agents, control sessions, review PRs. Anywhere.

  Dev tools · Sep 2026 · kilo.ai

• 
  AINA▲497

  Find your job search blind spots with an AI coach

  AI · Sep 2026 · aina-tech.io