AI Agent API Pricing Comparator

# Hermes Dispatch Tool — Agent API Pricing Comparator
# Source: https://hermesdispatch.dev/tools/agent-api-pricing-comparator/
# Description: Compare LLM API pricing across providers for a given token volume.
# License: MIT (generated by hermesdispatch.dev)
#
# INSTALL:
#   1. Save this file as ~/.hermes/hermes-agent/tools/agent_api_pricing_comparator.py
#   2. Restart Hermes or run /reset in a session
#   3. The tool auto-registers if Hermes uses auto-discovery of tools/*.py
#
# MANUAL REGISTRY (if auto-discovery is off):
#   from tools.agent_api_pricing_comparator import register
#   register()

import json

DATA = {"last_updated": "2026-06-29", "note": "Prices are per 1M tokens in USD. Check provider sites for current rates.", "providers": [{"name": "OpenAI GPT-4o", "model": "gpt-4o", "input_per_1m": 2.5, "output_per_1m": 10.0, "cached_input_per_1m": 1.25, "request_fee": 0.0, "context_window": 128000, "best_for": "General agent tasks, vision, tool use", "url": "https://openai.com/api/pricing/"}, {"name": "OpenAI GPT-4o-mini", "model": "gpt-4o-mini", "input_per_1m": 0.15, "output_per_1m": 0.6, "cached_input_per_1m": 0.075, "request_fee": 0.0, "context_window": 128000, "best_for": "Cheap high-volume agents", "url": "https://openai.com/api/pricing/"}, {"name": "Anthropic Claude 3.5 Sonnet", "model": "claude-3-5-sonnet-20240620", "input_per_1m": 3.0, "output_per_1m": 15.0, "cached_input_per_1m": 3.75, "request_fee": 0.0, "context_window": 200000, "best_for": "Long context, reasoning, coding", "url": "https://www.anthropic.com/pricing"}, {"name": "Anthropic Claude 3 Haiku", "model": "claude-3-haiku-20240307", "input_per_1m": 0.25, "output_per_1m": 1.25, "cached_input_per_1m": 0.03, "request_fee": 0.0, "context_window": 200000, "best_for": "Fast, cheap agent calls", "url": "https://www.anthropic.com/pricing"}, {"name": "Google Gemini 1.5 Pro", "model": "gemini-1.5-pro", "input_per_1m": 3.5, "output_per_1m": 10.5, "cached_input_per_1m": 0.0, "request_fee": 0.0, "context_window": 2000000, "best_for": "Massive context, multimodal", "url": "https://ai.google.dev/pricing"}, {"name": "Google Gemini 1.5 Flash", "model": "gemini-1.5-flash", "input_per_1m": 0.35, "output_per_1m": 1.05, "cached_input_per_1m": 0.0, "request_fee": 0.0, "context_window": 1000000, "best_for": "Low-latency, high-volume", "url": "https://ai.google.dev/pricing"}, {"name": "Groq Llama 3.1 70B", "model": "llama-3.1-70b-versatile", "input_per_1m": 0.59, "output_per_1m": 0.79, "cached_input_per_1m": 0.0, "request_fee": 0.0, "context_window": 131072, "best_for": "Fast open-weight inference", "url": "https://groq.com/pricing"}, {"name": "Together AI Llama 3.1 70B", "model": "llama-3.1-70b", "input_per_1m": 0.9, "output_per_1m": 0.9, "cached_input_per_1m": 0.0, "request_fee": 0.0, "context_window": 131072, "best_for": "Open-weight fine-tuning hub", "url": "https://www.together.ai/pricing"}, {"name": "Cohere Command R+", "model": "command-r-plus", "input_per_1m": 2.5, "output_per_1m": 10.0, "cached_input_per_1m": 0.0, "request_fee": 0.0, "context_window": 128000, "best_for": "RAG and enterprise search", "url": "https://cohere.com/pricing"}]}

def _ok(result):
    return json.dumps({"success": True, "data": result}, indent=2)

def _err(message):
    return json.dumps({"success": False, "error": message}, indent=2)


TOOL_NAME = "agent_api_pricing_comparator"
TOOLSET = "agents"

SCHEMA = {
  "type": "function",
  "function": {
    "name": "agent_api_pricing_comparator",
    "description": "Compare LLM API pricing across providers for a given token volume.",
    "parameters": {
      "type": "object",
      "properties": {
        "input_tokens": {
          "type": "integer",
          "description": "Input tokens per request."
        },
        "output_tokens": {
          "type": "integer",
          "description": "Output tokens per request."
        },
        "provider": {
          "type": "string",
          "description": "Optional provider name: OpenAI GPT-4o, OpenAI GPT-4o-mini, Anthropic Claude 3.5 Sonnet, Anthropic Claude 3 Haiku, Google Gemini 1.5 Pro, Google Gemini 1.5 Flash, Groq Llama 3.1 70B, Together AI Llama 3.1 70B, Cohere Command R+"
        }
      },
      "required": []
    }
  }
}

def _run(args):
    input_tokens = int(args.get("input_tokens", 1000))
    output_tokens = int(args.get("output_tokens", 500))
    provider_name = args.get("provider", None)
    providers = [p for p in DATA["providers"] if p["name"] == provider_name] if provider_name else DATA["providers"]
    results = []
    for p in providers:
        in_cost = (input_tokens / 1_000_000) * p.get("input_per_1m", 0)
        out_cost = (output_tokens / 1_000_000) * p.get("output_per_1m", 0)
        results.append({
            "provider": p["name"],
            "model": p.get("model", ""),
            "input_cost": round(in_cost, 5),
            "output_cost": round(out_cost, 5),
            "total_cost": round(in_cost + out_cost, 5),
        })
    results.sort(key=lambda x: x["total_cost"])
    return _ok({"results": results[:5], "input_tokens": input_tokens, "output_tokens": output_tokens})

def HANDLER(args):
    try:
        return _run(args)
    except Exception as e:
        return _err(str(e))


def register():
    """Manual registry hook. Import and call this to register with Hermes."""
    try:
        from tools.registry import registry
        registry.register(
            name=TOOL_NAME,
            toolset=TOOLSET,
            schema=SCHEMA,
            handler=HANDLER,
        )
    except ImportError:
        print("Hermes registry not found; skipping manual registration.")

if __name__ == "__main__":
    # CLI smoke test
    print(HANDLER({}))