OpenFill is in development. Inference is off, but is tested end to end, and will switch on at launch. All data currently on the site is for live testing: it will be erased at launch.

Skip to content
openfill

Model

market-preview

market-preview is its own market on OpenFill. A request names market-preview, sets a price limit, and pays the level the market is clearing at when it runs.

A preview of the market that opens once there is traction and compute behind it. Its prices are simulated until then; fixed-price models run today.

Open market

Datasheet

FieldValue
Model idmarket-preview
Also answers toNo other id
Context length262,144 tokens
Max output128,000 tokens
Pricing ratio1:10:20

The ratio weights a cache-hit, a cache-miss and an output token in the price. max_tokens above the max output is clamped down to it.

The market now

The clearing level, the price of each token kind at that level, and the quote a locked request bids.

Clearing level · market-preview · preview

$0.0395Basis−$0.0001 (−0.2%)Price lock$0.0394New price in 1s

Per 1M tokens: $0.0172 cache-hit · $0.1716 cache-miss · $0.3432 output

A price_lock request pays at most the locked price, and cancels with no charge if it has not started within 10 seconds.

The basis is the margin over the level, and a ceiling rather than a charge. See how the lock works.

Uptime, last 24 hours

Each check is a chat completion through the public API, run the way the models board describes.

Uptime

n/a

Checks

0

Failures

0

Latency

n/a

Queue wait

n/a

Checks begin when the market opens.

The first call

Set OPENFILL_API_KEY to a key from API keys before running an example. A key starts with of_live_ and is shown once.

Read docs
import osfrom openai import OpenAI client = OpenAI(    base_url=(        "https://api.openfill.ai/v1"    ),    api_key=os.environ[        "OPENFILL_API_KEY"    ],) r = client.chat.completions.create(    model="market-preview",    messages=[        {"role": "user", "content": "hi"}    ],    extra_body={"price_limit": 0.10},)print(r.choices[0].message.content)print(r.usage)

Questions

What does market-preview cost on OpenFill?

The clearing level when the request runs, up to the price limit you set, shared across the three token kinds by the ratio 1:10:20.

How pricing works

What happens while the level is above my price limit?

The request waits in the book while capacity is short and runs when there is room, for up to max_wait (10 seconds to 30 days). Past that it expires with 408 and no charge.

Waiting and timeouts

How do I know the price before I send?

With price_lock: true the request bids the published quote, or your price limit if lower, and starts within 10 seconds or cancels with 429 and no charge.

Price lock

What are the rate limits?

1,200 inference requests a minute per key, 2,000 queued orders per account, and a request body of up to 8 MB. Every rate limit is on GET /v1/rate-limits.

Rate limits

What is stored?

Storage is off by default: a prompt and its response exist only while the request runs. With it on, results are kept up to 30 days; store: false keeps less.

Data storage