available for preview

A8 Touchstone

Any criterion, in plain words. A verdict that clears the accuracy you set, or an abstention.

sample999916750585375
A touchstone. The sample’s streak is read against the streaks of needles of known fineness, in parts per thousand.

the stone

Gold had to prove its grade.

Theophrastus described the test around 300 BC. An assayer drew the gold across a black stone and left a streak. Beside it he drew needles of known fineness, one streak each, and read the gold’s color against theirs.

The needle it matched gave the grade. An unmatched streak was set aside.

999916sample750585375
The sample reads between 916 and 750.

the same five parts

A8 reads a subject the way an assayer read a streak.

The gold
The subject, sent as the user message.
The test
The criterion, in plain words, sent as the system message.
The needles
The examples you teach it.
The grade
The minimum accuracy, from 0 to 99, set on each call.
An unmatched streak
An abstention.

the grade

You set the grade. Below it, A8 abstains.

Every call carries the accuracy its verdict must meet, from 0 to 99. At 95, at least 95 answered verdicts in a hundred are correct. A case whose evidence falls short of 95 comes back as an abstention.

A8 finds the threshold for the grade itself. A labeled set, if you have one, tests the promise.

The verdict is scored from the evidence, so the same request returns the same verdict.

a call

Call it like a chat model. Set the grade on every call.

  1. 1

    The system message is the criterion: what you are judging.

  2. 2

    The user message is the subject: the thing under judgment.

  3. 3

    The schema is the verdict’s shape. A8 answers in JSON that matches it.

  4. 4

    The model’s tag sets the grade: a8:95 is 95, and 90 is the default. At 95, at least 95 answered verdicts in a hundred are correct. Below it, A8 abstains.

  5. 5

    Here A8 abstains, so a person decides. Tell your assistant their answer, and over MCP it teaches A8, once you approve.

from openai import OpenAI

client = OpenAI(
    base_url="https://api.u22a8.ai/v1",
    api_key="<your key>",
)

criterion = "Which queue does the ticket belong in?"  # 1
subject = "It arrived crushed. I want a refund."  # 2
messages = [
    {"role": "system", "content": criterion},
    {"role": "user", "content": subject},
]
queues = {"enum": ["billing", "shipping", "account"]}
route = {"type": "json_schema", "json_schema": {"name": "route",
    "schema": {"properties": {"queue": queues}}}}  # 3

resp = client.chat.completions.create(
    model="a8:95",  # 4: answer only at 95 or above
    messages=messages, response_format=route,
)
resp.choices[0].message.refusal
# 'Too close to call at min_accuracy 95; answered at 80.'

# 5: a person decides it's billing. Tell your assistant, over MCP:
#    "That ticket belongs in billing. Teach A8."

the needles

Teach it in one sentence.

another decision model

Rewording

  1. Reword the question until it answers right
  2. Re-run your labeled set
  3. Pick a new threshold

A8

One sentence

  1. Tell your assistant the answer you want
  2. Approve the teach it sends over MCP
  3. A new version fits in the background

The cases A8 abstains on are the ones to teach. A teach is kept the moment you approve it. About a minute after your workspace goes quiet, it trains a new version of your workspace’s model in the background, and the next verdicts come from it. Until then, the current version answers. The accuracy it promises is earned again on what you taught.

Ask why it answered, and it cites the examples you taught that decided it. Your examples stay in your workspace, and they train only your model.

the method, assayed

We tested the method against trained human raters.

On essay quality, we compared the scoring approach A8 is built on with trained human raters and with a panel of eight LLM judges. Its rank correlation with the raters was 0.815. The strongest judge on the panel reached 0.813.

read qed-bench
  1. u22a8 metric0.815
  2. claude opus 4.70.813
  3. claude sonnet 4.60.724
  4. deepseek v3.20.708
  5. llama 4 maverick0.687
  6. gemma 3 27b0.659
  7. mistral large 30.658
  8. qwen3 32b0.653
  9. claude haiku 4.50.637
Spearman rank correlation with trained human raters, essay quality (ASAP 2.0). Streaks start at zero.

at a glance

interfaces
OpenAI and Anthropic APIs, and MCP
minimum accuracy
0–99, set per call; A8 finds the threshold
determinism
same request, same verdict
verdicts
scored against the evidence
learning
a model per workspace, trained on what you teach
explanation
the taught examples behind a verdict
access
preview

read the model card

The stone is ours. The needles are yours.

State the criterion and the grade it must clear. Answer the cases A8 holds back on, and your workspace’s model learns from them. A8 is available for preview.

To build A8 into your stack or a tool, write to .

999916750585375yours