ModelsParticula-Code
Code that compiles,tests attached.
Particula-Code generates code that compiles and passes your tests before it answers. Every response ships with the tests it passed, in Python, TypeScript, Go, or Rust. It runs next to your repo, on your hardware, so proprietary code stays put.
YOUR-GPU · PARTICULA-CODE
$ POST /v1/generate · "parse EDI 850 orders"
→ particula-code · tests attached
→ "language": "python"
→ "tests_passed": "14 / 14" · "compiles": true
✓ 200 · 380 ms · tests green
ready
PARTICULA-CODE · YOUR HARDWARE · ZERO TELEMETRY
What it doesSee it on your data
Compiles before it answersA sandboxed loop compiles candidates and runs your test suite. What comes back is the best candidate that passed, not the first thing sampled.
Python, TypeScript, Go, RustTrained on the four languages, not eighty. Depth over coverage: idiomatic output in the stacks production teams actually run.
Tests attached to every responseEach answer carries the tests it passed, so review starts from evidence instead of a diff and a hope.
Repo-aware context64K tokens fits the files around the change. Point it at the module, not a snippet, and the output matches your conventions.
Spec sheetCompare all six
Context window64K tokensEnough for the surrounding module and its tests in one request.
Median latency380 ms medianBatch 1 on the reference single-GPU deployment. Our own measurement: the shipped harness reproduces it on your hardware.
InterfaceOpenAI-compatible RESTServed by vLLM, deployed with Docker Compose or Helm. Drop-in for existing OpenAI client code.
Hardware1× 48 GB GPUL40S or A6000 class. Sizing reviewed before the pilot.
Fine-tuningLoRA on your dataTuning jobs are included with Enterprise. A few hundred labeled examples from your pipeline is usually enough to move the needle.
How deployment worksStart at step 01
01Discovery call
30 minutes, no slides. Your task, your data shape, and the security constraints around it. If this model is the wrong fit, we say so on the call.02Pilot install
Containerized deployment to staging in your VPC or on-prem, on the same stack as Lumen and Notetaker. The eval harness runs on your data and the numbers go into the pilot agreement.03Production rollout
Runbook handoff, operator training, and 90 days on call. After that it is yours to operate, no vendor dependency.More modelsAll six models
Request a demo
“Bring a hundred real examples from your pipeline. We run the model on them live, the harness scores the output in front of you, and you decide with numbers instead of a sales deck.”
Sebastian Mondragon, Founder of Particula Tech
Request a demo30 minutes · no sales deck