Model Catalog
Every model routes through the Opsora gateway with automatic failover, load balancing, and usage tracking.
Coder
Code generation, completion, and review
qwen3-coder
Alibaba Qwen
Code generation & completion
deepseek-coder-v3
DeepSeek
Code generation & debugging
codestral-22b
Mistral AI
Code completion & review
starcoder2-15b
BigCode
Code completion across 600+ languages
codegemma-7b
Code understanding & generation
Ultra
Highest-capability reasoning and multimodal
nemotron-super
NVIDIA
General reasoning & analysis
gemini-2.5-pro
Multimodal reasoning, 1M+ context
gpt-4o
OpenAI
Multimodal general-purpose
claude-3.5-sonnet
Anthropic
Reasoning, coding & analysis
deepseek-v3
DeepSeek
Large-context reasoning
nemotron-ultra
NVIDIA
High-capability open reasoning
Fast
Low-latency, cost-effective inference
gemini-2.5-flash
Fast multimodal inference
llama-3.3-8b
Meta
Lightweight general-purpose
mistral-nemo
Mistral AI
Efficient multilingual (128k)
phi-3.5-mini
Microsoft
Compact reasoning (128k)
qwen2.5-7b
Alibaba Qwen
Efficient multilingual & math
Open
Open-weight models you can self-host
granite-3.1-8b
IBM
Enterprise open-weight
llama-3.3-70b
Meta
Large open-weight general
mistral-nemotron
Mistral / NVIDIA
Open reasoning & code
qwen2.5-coder-32b
Alibaba Qwen
Open code generation
All plans include access to every model. Pricing shown is per 1M tokens at the provider level.