ML Env

Use casesOperations & logistics

Find your SKU when the supplier uses their own product names and codes

Matches supplier, carrier and marketplace product descriptions to your own SKUs and suggests the top candidates for one-click confirmation.

How it works
Suppliers, carriers and marketplaces all describe the same product differently ('Bosch GSR 12V-15 FC Set' vs 'screwdriver bosch 12v fc 2x2.0Ah L-case'). Embeddings index your catalogue; when an unmatched description arrives as text — a line already read out of an order confirmation, invoice or delivery note — the system retrieves the closest SKUs, the reranker orders them, and a person sees the top three candidates to confirm with a click. Confirmed pairs are saved, so the same supplier string never needs matching twice. It is the unglamorous glue that lets extracted document lines actually land against the right SKU in your ERP.
Data you need
Your product catalogue export: SKU, names, key attributes, EAN/GTIN where you have them. Incoming descriptions must already be text — digital PDFs, CSV, EDI, email bodies. Scanned or photographed paperwork needs a separate OCR step first, as no vision model is served.
What to expect
Strong on distinct products; weakest on near-identical variants — same product, different pack size or colour code — which are exactly the cases where a wrong match is costly. Those need attribute rules (match on EAN or pack size) on top of similarity, and patchy EAN/GTIN coverage, common in SMB catalogues, weakens those rules. A sparse catalogue with bare titles and no attributes gives mediocre candidates.
Where people stay involved
A person confirms each new match; only previously confirmed pairs flow through automatically. Auto-accepting the top suggestion is how you book stock against the wrong SKU.

Which model, and what it costs to run

Qwen3-32B

This job answers while someone waits, so latency comes first: we hold it to a single small card and, within that, take the best independent score on structured output and tool calls.

Licence
Apache-2.0
Weights at 4-bit
18 GB
Context
32K tokens
Publisher
Alibaba (Qwen Team)

What the hardware costs

One 48 GB card holds it

Rent in the EU
$1.60/hrScaleway, Paris (PAR2)
Buy the card
$7,569new, one-off
Or rent it by the token
$0.12 / $0.12per M in / out · Nebius AI Studio · EU

Hardware only, third-party prices from 2026-07. The figure excludes the KV cache, which grows with context length and how many people use it at once — sized properly in a conversation, not guessed here. Renting by the token is cheaper up front; why our customers still self-host is below.

Structured output and tool calls

Berkeley Function Calling Leaderboard · v4 · 13 of 18 models measured

Measured in a sandbox, on somebody else's functions. It tells you which models are capable of the shape of the job, not which one will survive contact with your API.

GLM-4.672.38%Kimi K259.06%DeepSeek-V3.254.12%Qwen3-32B48.71%Qwen3-235B-A22B47.99%Qwen3-8B42.57%Qwen3-30B-A3B41.39%Qwen3-14B41.03%
Longer is betterFree to serveConditions apply⚠ answered under 95%

Independent measurement · UC Berkeley (Gorilla project) · board updated 2026-04-12

The API is cheaper per token. Here is why our customers don't use it.

We will not pretend otherwise: renting a model by the token from a serverless API costs less per million tokens than a card we run for you. We show that price on every use-case page. What it does not include is the part a shop with a customer database actually pays for.

  1. 01

    Your data never leaves hardware you can point at

    A serverless "we don't retain your data" is a clause in a contract. Running the model on a card in Amsterdam is a fact of architecture: your catalogue, tickets and customer records are never sent to a third party at all. For a GDPR audit, that is the difference between a promise and a floor plan.

  2. 02

    The price cannot move without your say-so

    A serverless rate card is somebody else's lever. The provider can raise the price, retire the model, or change the terms, and your cost moves with it. The same model on the same card costs the same next year — you own the number.

  3. 03

    The model cannot be taken away

    Hosted APIs deprecate models on their own schedule; the one you built on can be gone in a quarter. An open-weight model on your own hardware runs for as long as you keep the lights on. No vendor can end-of-life it out from under you.

And the price gap closes with volume: past a card you keep busy — very roughly four billion tokens a month — owning is cheaper outright, even before the three reasons above.

Getting a case like this one from a conversation to production takes about two months, and you can stop at the end of any phase.

Next

Tell us what your team does by hand

Describe the process that takes the most time. We will say plainly whether a model is the right tool for it.