HiNow LogoHiNow

Mercury 2

Model Information

Display Name: Mercury 2

API Model ID: inception/mercury-2

Category: Text To Text

Description: Mercury 2 is a reasoning diffusion LLM (dLLM) from Inception. Instead of generating tokens one after another, it produces and refines several tokens in parallel, which is where its speed comes from. **Key Features:** - 128,000-token context window - Up to 50K output tokens - Diffusion decoding: tokens generated and refined in parallel - Reasoning that can be turned on or off - Function/tool calling - Structured outputs (JSON schema) - Prompt caching for cheaper repeated context **Best For:** - Latency-sensitive assistants and interactive loops - High-volume reasoning where wall-clock time matters - Tool-driven workflows needing fast turnarounds **Note:** text input only — no image support.

Context Window: 128,000 tokens

Max Output: 50,000 tokens

How to Use This Model

To use Mercury 2 via the HInow.ai API, use the model ID: inception/mercury-2

API Request Example (Chat/Text)


POST https://api.hinow.ai/v1/chat/completions
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

{
  "model": "inception/mercury-2",
  "messages": [
    {"role": "user", "content": "Your message here"}
  ]
}

Pricing

  • input: $1.25
  • output: $3.75
  • cached: $0.125

Available Parameters

  • max_tokens: Max tokens to generate, including reasoning (Options: 512, 1024, 2048, 4096, 8192, 16384, 32768, 50000)
  • temperature: Sampling temperature
  • stop: Stop sequences
  • tools: Function/tool definitions for agentic workflows
  • tool_choice: Controls whether the model calls a tool (Options: auto, none, required)
  • response_format: Output format (structured outputs supported) (Options: text, json_object, json_schema)

Quick Reference

To use this model, set: "model": "inception/mercury-2"

Featured: No

Documentation: https://hinow.ai/models/inception/mercury-2

API Endpoint: https://api.hinow.ai/v1

Back to Models

Mercury 2

inception/mercury-2

$1.25 / $3.75
per 1M tokens (in/out)

About

Mercury 2 is a reasoning diffusion LLM (dLLM) from Inception. Instead of generating tokens one after another, it produces and refines several tokens in parallel, which is where its speed comes from.

Key Features:

  • 128,000-token context window
  • Up to 50K output tokens
  • Diffusion decoding: tokens generated and refined in parallel
  • Reasoning that can be turned on or off
  • Function/tool calling
  • Structured outputs (JSON schema)
  • Prompt caching for cheaper repeated context

Best For:

  • Latency-sensitive assistants and interactive loops
  • High-volume reasoning where wall-clock time matters
  • Tool-driven workflows needing fast turnarounds

Note: text input only — no image support.

Capabilities

Text To Text
Context128K tokens
Max Output50K tokens

Parameters

max_tokens

Max tokens to generate, including reasoning

5121024204840968192163843276850000
temperature

Sampling temperature

stop

Stop sequences

tools

Function/tool definitions for agentic workflows

tool_choice

Controls whether the model calls a tool

autononerequired
response_format

Output format (structured outputs supported)

textjson_objectjson_schema

Code Examples

curl -X POST https://api.hinow.ai/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $HINOW_API_KEY" \
  -d '{
    "model": "inception/mercury-2",
    "messages": [
      {"role": "user", "content": "Hello! How are you?"}
    ],
    "parameters": {
      "max_tokens": "512",
      "temperature": "",
      "stop": "",
      "tools": "",
      "tool_choice": "auto",
      "response_format": "text"
    }
  }'