AIVory Smart Inference

All Models. Live Pricing.

Browse every open-weight model in the Smart Inference catalogue. Each one routed to the cheapest live provider on every request.

DeepSeek V3.2 DeepSeek

671B params 160K context

code generation math reasoning multilingual

DeepSeek V4 Flash DeepSeek

MoE params 128K context

fast inference cost-effective general purpose

Gemma 2 9B Gemma

9B params 8K context

edge deployment classification extraction

Gemma 4 31B Gemma

31B params 128K context

instruction following lightweight deployment multilingual

Llama 3.3 70B Llama

70B params 128K context

general reasoning code generation instruction following

Llama 4 Maverick Llama

400B params 128K context

multimodal long context complex reasoning

Llama 4 Scout Llama

109B params 128K context

balanced cost-performance instruction following multilingual

Mistral Nemo 12B Mistral

12B params 128K context

efficient inference instruction following code

Mistral Small 24B Mistral

24B params 32K context

function calling structured output multilingual

Mistral Small 3.2 24B Mistral

24B params 128K context

vision function calling structured output

Mixtral 8x7B Mistral

46.7B params 32K context

cost-efficient MoE multilingual code

Qwen 2.5 72B Qwen

72B params 128K context

multilingual math code generation

Qwen 3 235B Qwen

235B params 128K context

frontier reasoning agentic workflows long-form content

Qwen 3.5 27B Qwen

27B params 128K context

agentic coding tool use instruction following

Voxtral Mini 3B Voxtral

3B params 32K context

voice input ultra-lightweight edge deployment

Voxtral Small 24B Voxtral

24B params 32K context

voice understanding multilingual speech transcription

Start using any model. Today.

One API key, every model, cheapest provider.

Credits from $10, no subscription, no commitment. Swap your base_url and ship.

Get early access Learn about Smart Inference