Mercury 3

inception/mercury-3

A chat model with a 1M-token context window, at $0.105 per million input tokens and $0.3675 per million output tokens.

Details

Maker: inception
Kind: Chat
Context window: 1,000,000 tokens
Max output: 131,072 tokens
Input: $0.105 per million tokens
Cached input: $0.0105 per million tokens, 90% off
Cache write: Same as input
Output: $0.3675 per million tokens
Capabilities: Reasoning, tool use
Takes: Text
Returns: Text
Zero data retention: Not on every provider
No training on requests: Yes, on every provider
Released: 2026-10-19

Description

Mercury 3 is Inception’s most capable model for agentic workflows and real-time applications. It is one of the world’s fastest models, powered by diffusion.

Call it

curl

curl https://phantom.codes/v1/chat/completions \
  -H "Authorization: Bearer $PHANTOM_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "inception/mercury-3",
    "messages": [
      {"role": "user", "content": "Hello"}
    ]
  }'

Set PHANTOM_KEY to your key. Any OpenAI client works with the base URL https://phantom.codes/v1.

Get a key Compare with another model