Y0 Text

available

The workhorse. Drafting, summarizing, and transforming language inside your context.

[ 01 ]Spec sheet

statusavailable
context window128k tokens
latency p50180 ms first token
price$0.40 in / $1.60 out

Y0 Text is the default surface of the cognitive runtime: the family you reach for when the job is language in, language out. It drafts the email, compresses the meeting into a brief, rewrites the proposal in your client's register, and extracts the four numbers that matter from a forty-page PDF. What separates it from a bare language model is the context graph underneath. A Y0 Text run does not start from a blank prompt — it starts from your documents, your calendar, and your prior runs, fetched through the trust kernel with an explicit scope, so the draft already knows the client, the deadline, and the tone you used last time. Two execution profiles cover the practical range: y0-fast answers interactively and is tuned for sub-200-millisecond first tokens; y0-deep trades latency for multi-pass composition on long or high-stakes documents. Both produce a replayable trace, so when a summary misses a clause you can see exactly which context item it read and which it never fetched. Y0 Text is generally available, serves every paid tier, and is the model behind most of the eight lanes' language surfaces.

[ 02 ]Capabilities

Drafting and rewriting in a persisted personal or brand voice

Long-document summarization with citation back to source spans

Structured extraction to JSON schemas you define per run

Translation and register shifts across 32 supported languages

Context-aware completion that reads the graph before it writes

[ 03 ]Best for

Inbox and document automation inside the work lane

Briefs, summaries, and reports generated from attached context

Schema-constrained extraction pipelines feeding other systems

[ 04 ]Sample request

[ request ]application/json
{
  "model": "y0-text-fast",
  "prompt": "Summarize the attached contract; flag renewal terms.",
  "context": ["documents"],
  "output_schema": { "type": "brief", "max_words": 250 },
  "max_steps": 2
}