Everything you need to run AI in your company.
Free tokens and hands-on help across three products: Australian-hosted models, a platform for all your AI agents, and an AI teammate in Slack and Teams.
$1,000 in free tokens for your first month. $2,000 for VC portfolio companies. First 100 companies.
- models
- Kimi K3, DeepSeek + 300 more
- agents
- TobyBot
- teammate
- Toby in Slack and Teams
- hosted
- Australia
Models, agents and a teammate. One team behind all three.
Start with the models. Add agents and Toby when your team is ready. Same company, same support, hosted in Australia.
$1,000 in free tokens
Spend them on the Model API in your first month. Our team helps you get agents and Toby running too.
Claim them→AU hosted models
Kimi K3 and DeepSeek V4.1 Flash hosted in Australia, plus 300+ more open models, or bring your own. OpenAI-compatible, so you change two lines of code.
Model API→A platform for all your agents
Build and run every kind of agent your team needs from one place, on our models or your own.
TobyBot→
An AI teammate in Slack and Teams
Toby lives in Slack and Microsoft Teams. Give it a task in a chat and it does the work. From $25 a month.
Meet Toby→For AI-native companies.
- AI is core to your product
- New to Xinference
- One of the first 100 companies to sign up
$2,000 in free tokens for every company you back.
Investing in AI-native companies? Bring us your portfolio. Each company gets double the standard amount, and our team helps them get set up on models, agents and Toby.
Join as a partner→From sign-up to first request.
- 01
Sign up
Your work email and company name. That’s it.
- 02
We set you up
Our team adds your free tokens and emails you the details.
- 03
Start building
Point your OpenAI SDK at the Model API. Add TobyBot agents and Toby when you’re ready.
client.py
client = OpenAI( base_url="https://api.openai.com/v1", api_key=OPENAI_API_KEY, base_url=XINFERENCE_BASE_URL, api_key=XINFERENCE_API_KEY,)
Two lines to switch. Your code stays.
- Chat completions and streaming
- Function calling and JSON mode
- TTFT and TPOT measured on every request
After your free tokens:
published per-token prices.
Hosted in Australia.
List prices as of October 2026, per 1M tokens. Billed per token.
Be ready when customers ask where the model runs.
Banks, insurers, hospitals and government buyers ask where their data goes before they sign.
Hosted in Australia
Kimi K3 and DeepSeek V4.1 Flash are hosted in Australia, and so is Toby.
No training on your data
We don’t use your prompts or outputs to train or fine-tune models unless you agree in writing.
Room to grow
Move a workload to dedicated GPUs or your own cloud account, on the same API.
Free token questions
What do we get?
$1,000 in free tokens on the Model API for your first month, plus help from our team to set up TobyBot agents and Toby for your company.
How many companies can join?
The first 100, in the order they sign up.
We’re a VC or accelerator. What do our companies get?
Each company you back gets $2,000 in free tokens. Join as a partner.
Do the free tokens cover Toby and TobyBot?
The free tokens are for the Model API. Toby and TobyBot have their own plans. Toby starts at $25 a month.
When do the free tokens expire?
At the end of your first month.
What happens when the free tokens run out?
You keep the same account and pay per token at published prices. Kimi K3 is $3.00 in and $15.00 out. DeepSeek V4.1 Flash is $0.30 in and $1.20 out, per 1M tokens. See pricing.
Which models can we use?
Kimi K3 and DeepSeek V4.1 Flash are ready on the Model API today. We offer 300+ more open models, or you can bring your own. Don’t see the model you need? Talk to us about adding it.
Is our data used for training?
No. We don’t use your prompts or outputs to train or fine-tune models unless you agree in writing.
Last updated
Get started
Claim your free tokens.
Just your work email and company name.
Our team adds your tokens and replies by email.
Switch your base URL, then add agents and Toby with our team.
Just want API access? Get API access

