ASI:One

The adaptive default model in the ASI:One family. Balances reasoning depth, speed, and cost - and activates the right capabilities based on your prompt.

---

Overview

asi1 is the adaptive default and a strong starting point for most workloads. It is designed for general agent workflows, chat, tool calling, and mixed workloads where requests vary in complexity from turn to turn.

If you are unsure which model to pick, start with asi1. Move to asi1-ultra when you need maximum reasoning depth and the largest tool budget, or to asi1-mini when latency and cost are the binding constraint.

---

Quickstart

cURL

curl -X POST https://api.asi1.ai/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $ASI_ONE_API_KEY" \
  -d '{
    "model": "asi1",
    "messages": [
      {"role": "user", "content": "Help me plan a trip to Tokyo"}
    ]
  }'

Python

import requests, os

url = "https://api.asi1.ai/v1/chat/completions"
headers = {
    "Content-Type": "application/json",
    "Authorization": f"Bearer {os.getenv('ASI_ONE_API_KEY')}"
}
data = {
    "model": "asi1",
    "messages": [
        {"role": "user", "content": "Help me plan a trip to Tokyo"}
    ]
}

response = requests.post(url, headers=headers, json=data)
print(response.json()["choices"][0]["message"]["content"])

JavaScript

import OpenAI from 'openai';
import dotenv from 'dotenv';

dotenv.config();

const client = new OpenAI({
  apiKey: process.env.ASI_ONE_API_KEY,
  baseURL: 'https://api.asi1.ai/v1',
});

const response = await client.chat.completions.create({
  model: 'asi1',
  messages: [
    { role: 'user', content: 'Help me plan a trip to Tokyo' },
  ],
});

console.log(response.choices[0].message.content);

---

Specifications

Specification	Value
Model ID	asi1
Position	Adaptive default
Context Window	200,000 tokens
Streaming	Supported
OpenAI Compatibility	Full SDK
APIs	Chat Completions (/v1/chat/completions), Responses (/v1/responses)
Rate Limits	See your API dashboard

asi1 shares the full 200,000-token context window with the rest of the family. The difference between models is in reasoning depth, tool budget, response size, and latency - not in what they can read.

---

Key Features

Adaptive Reasoning

Activates the right capabilities for each request. Light prompts return quickly; heavier ones get more depth.

Agentic by Default

Discovers and orchestrates agents from the Agentverse marketplace when a request benefits from specialized agents.

Full Tool Calling

Define your functions and the model decides when and how to call them. Full multi-turn and parallel tool support.

Web3 Native

Built for blockchain and decentralized applications. Understands smart contracts, tokenomics, and on-chain interactions.

---

What asi1 is for

General agent workflows

When your application has to handle a mix of request types - some short, some long, some that need tool calls, some that do not - asi1 adapts to each one. You do not have to route requests to different models manually.

Chat and tool calling

The default profile for interactive chat that occasionally needs to invoke tools, retrieve data, or call out to specialized agents. The right balance of latency and capability for everyday product flows.

Mixed workloads

When complexity varies from request to request and you do not want to maintain routing logic, asi1 is the safe default. Upgrade individual paths to asi1-ultra or downgrade them to asi1-mini once you have a signal.

Content creation, drafting, and editing

Multi-paragraph drafting, rewriting, and editing tasks where you want a thoughtful response without paying for the depth of asi1-ultra.

---

When to use a different model

If your workload requires...	Use this instead
Deep reasoning across many steps	asi1-ultra
Long agentic runs with many tool calls	asi1-ultra
Code review, contract analysis, deep research	asi1-ultra
Real-time chat, voice, autocomplete	asi1-mini
Classification or routing of short inputs	asi1-mini
High-volume, cost-sensitive workloads	asi1-mini

---

Tool calling with asi1

asi1 supports the full tool-calling API. Define your functions, pass them on every request, and the model decides when and how to call them.

import requests, os

url = "https://api.asi1.ai/v1/chat/completions"
headers = {
    "Content-Type": "application/json",
    "Authorization": f"Bearer {os.getenv('ASI_ONE_API_KEY')}"
}

tools = [
    {
        "type": "function",
        "function": {
            "name": "get_weather",
            "description": "Get the current weather for a city",
            "parameters": {
                "type": "object",
                "properties": {
                    "city": {"type": "string", "description": "City name"}
                },
                "required": ["city"],
            },
        },
    }
]

data = {
    "model": "asi1",
    "messages": [
        {"role": "user", "content": "What's the weather like in Tokyo right now?"}
    ],
    "tools": tools,
}

response = requests.post(url, headers=headers, json=data)
print(response.json())

For schemas, multi-turn flows, and parallel tool calls, see the Function Calling guide.

---

Trust & Transparency

Capability	Detail
Reasoning Traces	Returns reasoning steps with clear provenance
Reduced Hallucination	Confidence thresholds and cross-checks keep outputs grounded
Safety Filters	Built-in guardrails screen for harmful content
Audit Support	Hashed reasoning steps available for compliance logging

---

Related Models

• asi1-ultra - Deepest reasoning, longest agentic runs, up to 500 tool calls per turn
• asi1-mini - Fastest response, lowest latency, lightest cost

---

Ready to get started? Check out the Quick Start guide or explore OpenAI compatibility for easy integration.