Seedance 2.0 is here - create consistent, multimodal AI videos faster with images, videos, and audio in one prompt.

Try Now
Skip to main content
Available now on ModelsLab · Language Model

NousResearch: Hermes 2 Pro - Llama-3 8BFunction Calling Perfected

Try NousResearch: Hermes 2 Pro - Llama-3 8BAPI Documentation

Master Tools and JSON

Core Upgrade

Retrain on OpenHermes 2.5

Uses cleaned OpenHermes 2.5 dataset plus in-house function calling data for reliable outputs.

90% Accuracy

Excels at Function Calls

Scores 90% on Fireworks.AI eval; supports multi-turn with ChatML and special tokens like <tools>.

JSON Mastery

Structured Outputs

Achieves 84% on JSON eval; enables agentic parsing during streaming with single-token tags.

Examples

See what NousResearch: Hermes 2 Pro - Llama-3 8B can create

Copy any prompt below and try it yourself in the playground.

JSON Weather Query

“<|im_start|>system You are a helpful assistant with access to tools. Use JSON for structured responses.<|im_end|> <|im_start|>user Get current weather in Tokyo using tool.<|im_end|> <|im_start|>assistant <tools><tool_call id="1"><name>get_weather</name><parameters>{"city":"Tokyo"}</parameters></tool_call></tools><|im_end|>”

Multi-Turn Calc

“<|im_start|>system Handle math via calculator tool. Respond in JSON.<|im_end|> <|im_start|>user Calculate 25 * 4 + 17.<|im_end|> <|im_start|>assistant <tool_call id="calc"><parameters>{"expr":"25*4+17"}</parameters></tool_call><|im_end|>”

Code Gen Task

“<|im_start|>system Generate Python code. Output as JSON with reasoning.<|im_end|> <|im_start|>user Write function to sort list ascending.<|im_end|> <|im_start|>assistant <json>{"reasoning":"Use sorted() builtin.", "code":"def sort_list(lst): return sorted(lst)"}</json><|im_end|>”

Agent Workflow

“<|im_start|>system Parse tools for agent tasks. Use <tool_response> after calls.<|im_end|> <|im_start|>user Book flight to NYC on tool.<|im_end|> <|im_start|>assistant <tools><tool_call id="book_flight"><name>search_flights</name><parameters>{"dest":"NYC"}</parameters></tool_call></tools><|im_end|>”

For Developers

A few lines of code.
JSON Tools. ChatML Format.

ModelsLab handles the infrastructure: fast inference, auto-scaling, and a developer-friendly API. No GPU management needed.

  • Serverless: scales to zero, scales to millions
  • Pay per token, no minimums
  • Python and JavaScript SDKs, plus REST API
API Documentation 
import requests


response = requests.post(
    "https://modelslab.com/api/v7/llm/chat/completions",
    json={
  "key": "YOUR_API_KEY",
  "prompt": "",
  "model_id": ""
}
)
print(response.json())

FAQ

Common questions about NousResearch: Hermes 2 Pro - Llama-3 8B

Read the docs

Ready to create?

Start generating with NousResearch: Hermes 2 Pro - Llama-3 8B on ModelsLab.

Try NousResearch: Hermes 2 Pro - Llama-3 8BAPI Documentation