Intelligence
Models
Every model AVINCO can route to. AVINCO IQ runs built-in with no key; other models run once their provider is connected, and are routed through IQ until then.
Default model
AVINCO IQ Fast
Used for new chats and missions
Available now
2 / 15
Models with a ready provider
Routing
Automatic fallback
Unconnected models fall back to IQ
AVINCO IQ Fast
AVINCO IQ · 1M context
Low-latency everyday intelligence for chat, triage and routing.
AVINCO IQ Pro
AVINCO IQ · 1M context
Deeper reasoning for mission planning, architecture and code review.
GPT-4.1
OpenAI · 1M context
Flagship OpenAI model with strong instruction following.
o4-mini
OpenAI · 200K context
Compact reasoning model for math and planning.
Claude Sonnet
Anthropic · 200K context
Balanced Claude model for agentic coding and analysis.
Claude Haiku
Anthropic · 200K context
Fast, affordable Claude for high-volume tasks.
Gemini 2.5 Pro
Google AI · 1M context
Long-context multimodal reasoning.
Gemini 2.5 Flash
Google AI · 1M context
Fast multimodal model with a generous free tier.
Llama 3.3 70B
Groq · 128K context
Open-weight Llama served at extreme speed.
DeepSeek R1
DeepSeek · 128K context
Open reasoning model with visible chain-of-thought.
Mistral Large
Mistral · 128K context
Top-tier Mistral model, strong in EU languages.
Qwen 2.5 Coder 32B
Ollama · 32K context
Strong local coding model, runs offline.
Gemma 3 12B
LM Studio · 128K context
Compact Google open model for local use.
Nemotron 3 Super 120B
Custom Endpoint · 128K context
NVIDIA's large Nemotron-3 model, served through the NIM OpenAI-compatible endpoint.