Canadian AI infrastructure.
Built for developers.
Inference API with OpenAI and Anthropic SDK compatibility. GPU compute for training and fine-tuning. macOS environments for Claude Code and agents. All hosted in Canada, all transparent pricing.
const client = new Anthropic({
apiKey: process.env.RUNARS_API_KEY,
baseURL: "https://runars.ca/v1",
})Match your workload to the right compute.
Integrate a model, chat with one, run your own GPU workloads, or deploy full macOS environments for agents — all on one platform.
3.0 Products →
Inference API
Drop-in OpenAI and Anthropic-compatible API endpoints. Run models like GLM-5.3 Flash with your existing SDK without changing a line.
ANTHROPIC_BASE_URL and OPENAI_BASE_URLGPU Cloud
Rent GPU compute on-demand. Serverless for bursts, dedicated for sustained workloads. Bring your own container.
Cloud MacFuture
On-demand macOS VMs for Claude Code and Codex sessions. Spin up a fully provisioned agent environment in seconds, integrate with any automation workflow.
An assistant that shows its work.
Runars Chat answers with the tools on display — live web search, CanLII legal research, and your own documents. Every tool call and every source stays visible, hosted in Canada.
See our live catalog rates.
Browse available model pricing and inspect real-time rates across our catalog. No hidden fees, no surprise overages.
| Model | Country | Input ($/M) | Output ($/M) | Context |
|---|---|---|---|---|
| GLM 5.3 Flash | 🇨🇦 | $0.2424/M | $0.8080/M | 1M |
| Deepseek V4 Flash | 🇨🇦 | $0.2644/M | $1.0577/M | 1M |
| Minimax M2.7 | 🇨🇦 | $0.5729/M | $2.2918/M | 198K |
| Qwen 3.6 35B A3B | 🇨🇦 | $0.4731/M | $2.8327/M | 262K |
| Deepseek V4 Pro 0813 | 🇨🇦 | $0.9696/M | $2.9088/M | 1M |
| Deepseek V4 Pro | 🇨🇦 | $1.2605/M | $3.7814/M | 1M |
| Aion Labs Aion 2.0 | 🇨🇦 | $1.9398/M | $3.8795/M | 128K |
| GLM 5 | 🇨🇦 | $1.6160/M | $5.1712/M | 198K |
| GLM 5.2 | 🇨🇦 | $1.6968/M | $5.3328/M | 1M |
| Qwen 3.6 27B | 🇨🇦 | $1.1459/M | $6.8753/M | 256K |
| GLM 5.1 | 🇨🇦 | $2.2624/M | $7.1104/M | 200K |
| GLM 5.3 | 🇨🇦 | $2.2624/M | $7.1104/M | 1M |
| Qwen 3.8 Max | 🇨🇦 | $4.7688/M | $14.3065/M | 1M |
| Kimi K3 | 🇨🇦 | $5.7294/M | $28.6472/M | 1M |
Pricing shown in CAD. Billing by token. All prices subject to change — check docs for current rates.
Built for Canada.
Your data doesn't leave Canadian territory. Not in flight, not at rest, not in backups.
4.0 Infrastructure →
Montreal & Calgary
Compute runs on our own Canadian GPU Cloud across two Canadian data centres. Low latency coast-to-coast. All Canadian.
No data residency surprises
Your inference requests run in Canada. Your prompt cache stores in-region. Your logs stay on-shore.
For regulated industries
Built for financial services, healthcare, and government workloads where data sovereignty is non-negotiable.
Why Canadian infrastructure matters
For regulated sectors—financial services, healthcare, government research—Canadian data residency is the policy, not the feature. We built Runars to be that policy's infrastructure.
Turn capacity into revenue.
Have unused inference capacity? Tell us about it. We're always looking to expand our provider network with reliable capacity partners.
Capacity Partners →
Share available capacity
Tell us where you have available inference capacity and its approximate dollar value. Include your provider type and monthly capacity estimate.
We review provider fit
Our infrastructure team evaluates the technical fit, pricing alignment, and reliability requirements. We'll verify performance SLAs and capacity.
Discuss supply and settlement
Once approved, we'll work with you on pricing, settlement terms, and integration. Most partners start their first workload within 2 weeks.
Get in touch
Fill out the form and our team will review your capacity fit and contact you within 2 business days with next steps.