◦ CATALOG

Models.

Open-source models running on sovereign Canadian infrastructure. 100% data residency. OpenAI-compatible endpoints.

GPT-OSS 120B

ONLINELLM

gpt-oss-120b

Largest open-source GPT model. Best quality for complex tasks, analysis, and generation.

PARAMS
120B
CONTEXT
131K
PROVIDER
OpenAI
PRICING
$0.11 in / $0.59 out per M
from openai import OpenAI

client = OpenAI(
    base_url="https://api.qubittron.ai/v1",
    api_key="qbt_YOUR_KEY"
)

response = client.chat.completions.create(
    model="gpt-oss-120b",
    messages=[{"role": "user", "content": "Hello"}]
)
print(response.choices[0].message.content)

Llama 3.3 70B Instruct

ONLINELLM

llama-70b

High-quality model with strong instruction following and reasoning capabilities.

PARAMS
70B
CONTEXT
131K
PROVIDER
Meta
PRICING
$0.93 in / $0.93 out per M

Qwen 3 Coder 30B

ONLINECODE

qwen-coder

Purpose-built for code generation. 262k context for entire codebases. MoE architecture.

PARAMS
30B (3B active)
CONTEXT
262K
PROVIDER
Qwen
PRICING
$0.09 in / $0.33 out per M

DeepSeek R1 Distill 70B

ONLINEREASONING

deepseek

Chain-of-thought reasoning model. Strong at math, science, and complex problem solving.

PARAMS
70B
CONTEXT
131K
PROVIDER
DeepSeek
PRICING
$0.93 in / $0.93 out per M

Mistral Small 3.2 24B

ONLINELLM

mistral-small

Latest Mistral small model. Great quality in a compact size with 131k context.

PARAMS
24B
CONTEXT
131K
PROVIDER
Mistral AI
PRICING
$0.13 in / $0.39 out per M

Qwen 2.5 VL 72B

ONLINEVISION

qwen-vl

Multimodal vision-language model. Understands images and text together.

PARAMS
72B
CONTEXT
32K
PROVIDER
Qwen
PRICING
$1.26 in / $1.26 out per M