Skip to content
YachayCondor Models
ModelsPricingDocs

Docs

How Yachay works

Public docs are being filled in alongside v1. The pages marked “Live” are ready now; the rest will land before launch.

Get started

What is Yachay

Live

A 5-minute tour of what gets fine-tuned, how, and why.

Model catalog

Live

All 17 base models, license terms, typical tune cost.

Pricing

Live

From ~$0.15 per million training tokens on production-size jobs — cheaper than Fireworks and Together. $5 minimum per job, no upfront bond — card on file, charged at completion.

Building a tune

Dataset format

Live

JSONL (OpenAI chat / Alpaca / ShareGPT) or CSV / TSV — auto-detected and normalized. Linter runs at submit.

Picking a base model

Live

Size vs. license vs. tune cost decision matrix.

LoRA vs QLoRA

Live

When to use each, and why we pick by default.

After the tune

Downloading your adapter

Live

30-day download window via fresh 1-hour signed URLs, paid extension at $0.06/GB-month + $0.50 handling (24 months max).

Eval reports

Draft

What metrics ship today (val perplexity + training loss in logs), and the v1.1 dashboard panel + benchmark-probe roadmap.

API access

Draft

Endpoints the dashboard uses today (first-party only), plus the v1.1 public-API + CLI roadmap.

Need answers before docs ship? Email hello@condorbox.ai — we’ll send the relevant draft directly.

YachayCondor Models

Custom AI fine-tuning. Fine-tune open-source LLMs on your own data, in the cloud — no GPU, no GCP account required.

Product

  • Model catalog
  • Pricing
  • Documentation
  • FAQ
  • Dashboard

Company

  • condorbox.ai
  • Status
  • Terms
  • Privacy
  • Contact
© 2026 Chicha Technology LLC. Yachay is Quechua for knowledge.