Supported Models

Convoy supports a wide range of models. Use the Convoy Model ID in your requests — Convoy handles routing to the underlying provider automatically.

Building or testing an integration? Use convoy-mock — a free, synthetic model that returns a callback in ~60 seconds and is never billed. It runs the entire batch pipeline so you can verify your wiring without burning real credits, then swap to a production model below when you’re ready.

Testing

convoy-mock — free, synthetic, ~60s callback round-trip — use this for all integration testing

Browse by Provider

Anthropic — Claude 3, 3.5, 3.7, Sonnet 4 / 4.5 / 4.6, Opus 4.5 / 4.6, Haiku 4.5, plus extended-context (*-long) variants
Amazon Nova — Nova Micro, Lite, Pro, Premier, Nova 2 Lite, Nova 2 Pro / Omni (Preview), and Titan Text Embeddings v2
Meta Llama — Llama 3.1, 3.2, 3.3, Llama 4 Maverick / Scout
Mistral AI — Mistral Large 2407 / 3, Devstral 2, Magistral, Ministral 3, Voxtral
DeepSeek — DeepSeek V3.1, V3.2
Google — Gemma 3 (4B, 12B, 27B)
MiniMax — MiniMax M2, M2.1
NVIDIA — Nemotron Nano 9B v2, 12B v2, Nano 3 30B
Moonshot AI — Kimi K2 Thinking, K2.5
OpenAI — GPT OSS 20B / 120B, GPT OSS Safeguard 20B / 120B
Qwen — Qwen3 32B, 235B, Coder, Coder Next, Next 80B, VL
Z.AI — GLM 4.7, GLM 4.7 Flash

Example Request


curl -X POST https://api.cnvy.ai/cargo/load \
  -H "Content-Type: application/json" \
  -H "X-API-Key: convoy_sk_your_key_here" \
  -d '{
    "params": {
      "model": "claude-3-haiku"
    },
    "records": [
      {
        "recordId": "rec_001",
        "modelInput": {
          "max_tokens": 1024,
          "messages": [
            {
              "role": "user",
              "content": "Summarize this document."
            }
          ]
        }
      }
    ]
  }'