Standard One

Send text or an image,
get a decision back

Classification for the calls your product makes all day: spam or not, which label, yes or no, how urgent. It answers fast, with one of your own options, and says how sure it is.

From $0.018/1M toks, 53% off. Start with free credit.

standard-one-3b

Anywhere your product has to make a call.

Product photos, parcels on a belt and moves in a game work the same way: you write one question and the options, and it returns a probability for each option. Pick One 8B where a wrong call costs the most and One 3B where speed matters most.

Game moves from text

One 3B2–4 options

Each request describes the game state as text, such as health, ammo and where the nearest enemy is. One 3B picks the next move from two to four options, 316 times across four scenes of Doom.

Highlights from real-time runs. The model reads the game state as text, not the screen.

Parcel inspection

One 8B2 options

One photo of each parcel goes to One 8B with two options: pass, or review for damage. Of 30 parcels it passed 28 and sent 2 to review.

AI-generated conveyor footage at 2.5× speed. Capture points and image regions set by hand. Real model outputs.

Catalog photo review

One 3B3 options

Each product photo goes to One 3B with one question, whether it suits a product-only catalog, and three options: keep, review or reject. Of these 120 photos it kept 93, sent 11 to review and rejected 16.

AI-generated sample photos. Real model outputs, edited playback.

Scores on screen are the model’s option probabilities, not accuracy. Measured results are in the evaluation.

Not a chatbot. An answer your code can use.

A general-purpose model writes a paragraph that someone, or some code, still has to read.
Standard One returns one of your options and how sure it is.

Noor Haddad10:31

Delivered to the wrong porch

Tracking says my order #57120 was delivered at 10:14, but the delivery photo shows a porch that isn’t mine. I need it before I fly on Friday.

delivery-photo.jpg

General-purpose model3,481 ms

It sounds like the carrier may have left order #57120 at the wrong address, since the proof-of-delivery photo shows a different porch. Misdeliveries like this are fairly common, and the parcel often turns up within a day or two. The customer could ask neighbours and the carrier, and it’s also worth checking the address saved on the account. If it can’t be found, a replacement or a refund may be appropriate, depending on your policy. Given the Friday flight, this seems fairly time-sensitive. If you’d like, I can draft a reply to the customer.

…and your code still has to parse it.

One 3B
22.6 ms
RouteShipping 0.94
ReshipYes 0.86
UrgencyHigh 0.71

Your code acts above 0.85Route to Shipping · Reship the order

A person decidesUrgency

Illustrative output comparison; animation timings are not a head-to-head benchmark. How thresholds work.

Scored on a public benchmark, beside much larger models.

On the Decision Index 0.3 public index, One 8B SH scores 57.4, within a point of the hosted Jev 1.13. One 3B SH scores 51.5.

Public index against model size

Chance-corrected: 0 = random guessing, 100 = perfect. Board data of October 12, 2026. One 8B SH v4 and One 3B SH v3, the versions the API serves, are self-scored with the official kit; board submission under review.

Pick the size that fits each decision.

Both take the same request, read text and images, and return the same kind of answer.

Standard One 3B

standard-one-3b

Faster, for high-volume calls.

Realistic decisions
88.2%
Median latency
22.6 ms
Price
$0.018 per 1M text tokens$0.055 per 1M image tokens
Inputs
TextImage
Availability
API · Open weights (Apache 2.0)

Best for spam and phishing filters, topic labels and other high-volume checks.

Weights on Hugging Face

Standard One 8B

standard-one-8b

Stronger on hard cases, for high-stakes calls.

Realistic decisions
91.5%
Median latency
25.8 ms
Price
$0.035 per 1M text tokens$0.08 per 1M image tokens
Inputs
TextImage
Availability
API · Open weights (Apache 2.0)

Best for agent routing, answer checks and intent across languages.

Weights on Hugging Face