Amos 1 mini

A 0.5B decision model. Give it a case (text, JSON or a photo) and a schema of typed questions; it answers every question in one forward pass, each with a probability for every answer you allow. It writes no text, so there is nothing to parse and no malformed output. Small enough to run on a laptop CPU, with no GPU: about 93 ms a request. The hosted API's default.

Part of the Amos 1 family: large · medium · small · mini. Playground · API docs · Model page

Results

Typed decisions, test split: 2,000 decisions from four workflows it never trained on, zero-shot.

	Accuracy	KL ↓	Brier ↓
Amos 1 mini	64.6%	0.225	0.121
Microsoft-Decision-1 (closed weights)	79.5%	0.318
Decider 1 (its makers' number)	76.8%	0.096
Perplexity Decider v1.1	73.9%	0.466
Clef (Cloudflare)	72.5%	0.191

Rivals run by us on the same 2,000 decisions, except Decider 1, whose number is its makers' own.

With each question's priors fitted on the benchmark's own train split and applied at answer time (the weights do not change), as Decider 1 also reports: 70.25% (Decider 1: 80.1%).

Decision Index 0.3, public part, with the board's own kit: 17.35. The board's ranking adds private tests that only its maintainers run.

Other benchmarks, as served
JevBench public easy / standard / hard	100.0 / 77.8 / 53.2
Ten public benchmarks (RAGTruth, JudgeBench, Belebele, WinoGrande, ContractNLI, TruthfulQA, Circa, Financial PhraseBank, TabFact, BBH)	65.0%
AG News / DAIR Emotion / Banking77 (77 labels)	81.5 / 55.6 / 80.5
Invoice totals quoted exactly	9 / 10
Points on the named object, RefCOCO	50%
Points on the named object, COCO	89%
Pet breeds	93.6%
Server time, median, three-question request	27 ms

Use

Hosted, no key: POST https://api.amos.scira.ai/v1/systemone with "model": "amos-1-mini".

Or run it yourself. The serving package is in serve/ (Python 3.12, PyTorch 2.14; versions pinned to the ones these numbers were measured with):

hf download mukaddamzaid/amos-1-mini --local-dir amos-1-mini
pip install ./amos-1-mini/serve
python -m amos.serve amos-1-mini --device cpu --port 8000
curl http://127.0.0.1:8000/v1/systemone -H 'Content-Type: application/json' -d '{
  "state": "Invoice INV-2231 for £306.00 was charged to my card twice. Please refund one.",
  "questions": {
    "refund": {"type": "noul", "instructions": "Is the customer asking for a refund?"},
    "team": {"type": "choice", "instructions": "Which team should handle this?",
             "criteria": {"billing": null, "technical": null, "sales": null}}
  }
}'

Question types: yes/no (noul), pick one (choice), rate (score), pick any (multi), quote (span) and point (region, with a photo in images). It runs on a laptop CPU and reads up to 9,216 tokens per request.