New Flagship Model Added: GPT-5.5

Date: 2026-04-25 / (1405-02-05)

Summary

OpenAI's newest flagship model GPT-5.5 is now available on AvalAI through both the Responses API (v1/responses) and Chat Completions API (v1/chat/completions). GPT-5.5 delivers state-of-the-art performance across agentic coding, knowledge work, computer use, and scientific research while maintaining per-token latency comparable to GPT-5.4.

---

Details

OpenAI

We announce access to GPT-5.5 (gpt-5.5), OpenAI's smartest and most intuitive model yet, designed for real work across complex, multi-step tasks. Documentation

GPT-5.5

GPT-5.5 is OpenAI's newest flagship model, engineered to carry more of the work itself: planning, using tools, checking its own output, navigating ambiguity, and persisting across long-running tasks. It excels at writing and debugging code, researching online, analyzing data, creating documents and spreadsheets, operating software, and moving between tools until a task is finished.

Key Features:

• Context Window: 1,000,000 input tokens, 128,000 max output tokens
• Input/Output: Text and image input, text output
• Advanced Reasoning: Configurable reasoning effort (none, low, medium, high, xhigh)
• State-of-the-Art Performance: 82.7% Terminal-Bench 2.0, 73.1% Expert-SWE, 84.9% GDPval, 78.7% OSWorld-Verified, 98.0% Tau2-bench Telecom
• Agentic Coding Excellence: Stronger long-horizon reasoning, context retention across large codebases, and more reliable tool use
• Token Efficiency: Uses significantly fewer tokens than GPT-5.4 to complete the same Codex tasks
• Endpoint Support: Available on v1/responses and v1/chat/completions

Pricing Details:

Model	Input	Cached Input	Output
gpt-5.5	$5.00/1M tokens	$0.50/1M tokens	$30.00/1M tokens

---

API Request/Response Examples

Chat Completions Example Request

curl https://api.avalai.ir/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AVALAI_API_KEY" \
  -d '{
    "model": "gpt-5.5",
    "messages": [
      {
        "role": "user",
        "content": "Refactor this Python function into an idiomatic async version and explain the trade-offs."
      }
    ]
  }'

Example Response

{
  "id": "chatcmpl-abc789",
  "created": 1777881600,
  "model": "gpt-5.5",
  "object": "chat.completion",
  "system_fingerprint": null,
  "choices": [
    {
      "finish_reason": "stop",
      "index": 0,
      "message": {
        "content": "Here's an idiomatic async rewrite using asyncio.gather for concurrent I/O, with explicit cancellation handling and a concise explanation of the trade-offs compared to the synchronous version...",
        "role": "assistant",
        "thinking_blocks": [],
        "annotations": []
      }
    }
  ],
  "usage": {
    "completion_tokens": 412,
    "prompt_tokens": 28,
    "total_tokens": 440,
    "completion_tokens_details": null,
    "prompt_tokens_details": {
      "audio_tokens": null,
      "cached_tokens": null,
      "text_tokens": 28,
      "image_tokens": null
    }
  },
  "estimated_cost": {
    "unit": "0.0124000000",
    "irt": 1420.8,
    "exchange_rate": 114600
  }
}

Responses API Example Request

curl https://api.avalai.ir/v1/responses \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AVALAI_API_KEY" \
  -d '{
    "model": "gpt-5.5",
    "input": "Analyze this quarterly financial report and surface the three most important risk factors with supporting evidence.",
    "reasoning": {
      "effort": "high"
    }
  }'

---

SDK Usage Examples

Bash

curl https://api.avalai.ir/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AVALAI_API_KEY" \
  -d '{
    "model": "gpt-5.5",
    "messages": [
      {
        "role": "user",
        "content": "Design a distributed rate-limiter that survives partial node failures."
      }
    ]
  }'

Python

from openai import OpenAI

client = OpenAI(api_key="your-avalai-api-key", base_url="https://api.avalai.ir/v1")

completion = client.chat.completions.create(
    model="gpt-5.5",
    messages=[
        {
            "role": "user",
            "content": "Design a distributed rate-limiter that survives partial node failures.",
        }
    ],
)

print(completion.choices[0].message.content)

JavaScript

import { OpenAI } from "openai";

const client = new OpenAI({
  apiKey: process.env.AVALAI_API_KEY,
  baseURL: "https://api.avalai.ir/v1",
});

const completion = await client.chat.completions.create({
  model: "gpt-5.5",
  messages: [
    {
      role: "user",
      content: "Design a distributed rate-limiter that survives partial node failures.",
    },
  ],
});

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

---

Advanced Features

Reasoning Effort Control

GPT-5.5 supports configurable reasoning effort so you can balance latency and depth on a per-request basis:

Bash

curl https://api.avalai.ir/v1/responses \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AVALAI_API_KEY" \
  -d '{
    "model": "gpt-5.5",
    "input": "Prove, step by step, that the sum of two odd integers is even.",
    "reasoning": {
      "effort": "high"
    }
  }'

Python

response = client.responses.create(
    model="gpt-5.5",
    input="Prove, step by step, that the sum of two odd integers is even.",
    reasoning={"effort": "high"},
)

print(response.output)

JavaScript

const response = await client.responses.create({
    model: "gpt-5.5",
    input: "Prove, step by step, that the sum of two odd integers is even.",
    reasoning: { effort: "high" },
});

console.log(response.output);

Function Calling

GPT-5.5 brings stronger tool use, staying on task for longer and selecting tools more reliably in agentic workflows:

Bash

curl https://api.avalai.ir/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AVALAI_API_KEY" \
  -d '{
    "model": "gpt-5.5",
    "messages": [
      {
        "role": "user",
        "content": "What is the weather like in Tokyo and should I pack a jacket?"
      }
    ],
    "tools": [
      {
        "type": "function",
        "function": {
          "name": "get_weather",
          "description": "Get current weather information for a location",
          "parameters": {
            "type": "object",
            "properties": {
              "location": {
                "type": "string",
                "description": "City name"
              },
              "unit": {
                "type": "string",
                "enum": ["celsius", "fahrenheit"],
                "description": "Temperature unit"
              }
            },
            "required": ["location"]
          }
        }
      }
    ],
    "tool_choice": "auto"
  }'

Python

tools = [
    {
        "type": "function",
        "function": {
            "name": "get_weather",
            "description": "Get current weather information for a location",
            "parameters": {
                "type": "object",
                "properties": {
                    "location": {
                        "type": "string",
                        "description": "City name",
                    },
                    "unit": {
                        "type": "string",
                        "enum": ["celsius", "fahrenheit"],
                        "description": "Temperature unit",
                    },
                },
                "required": ["location"],
            },
        },
    }
]

response = client.chat.completions.create(
    model="gpt-5.5",
    messages=[
        {
            "role": "user",
            "content": "What is the weather like in Tokyo and should I pack a jacket?",
        }
    ],
    tools=tools,
    tool_choice="auto",
)

JavaScript

const tools = [
    {
        type: "function",
        function: {
            name: "get_weather",
            description: "Get current weather information for a location",
            parameters: {
                type: "object",
                properties: {
                    location: {
                        type: "string",
                        description: "City name",
                    },
                    unit: {
                        type: "string",
                        enum: ["celsius", "fahrenheit"],
                        description: "Temperature unit",
                    }
                },
                required: ["location"],
            },
        },
    }
];

const response = await client.chat.completions.create({
    model: "gpt-5.5",
    messages: [{role: "user", content: "What is the weather like in Tokyo and should I pack a jacket?"}],
    tools: tools,
    tool_choice: "auto",
});

---

Benchmark Highlights

GPT-5.5 reaches new state-of-the-art results across coding, knowledge work, and tool use:

Benchmark	GPT-5.5	GPT-5.4
Terminal-Bench 2.0	82.7%	75.1%
Expert-SWE (Internal)	73.1%	68.5%
GDPval (wins or ties)	84.9%	83.0%
OSWorld-Verified	78.7%	75.0%
Tau2-bench Telecom	98.0%	92.8%
Toolathlon	55.6%	54.6%
BrowseComp	84.4%	82.7%
FrontierMath Tier 1–3	51.7%	47.6%
FrontierMath Tier 4	35.4%	27.1%
CyberGym	81.8%	79.0%
ARC-AGI-2 (Verified)	85.0%	73.3%

GPT-5.5 also delivers strong long-context retrieval, scoring 74.0% on OpenAI MRCR v2 (8-needle, 512K–1M) compared to 36.6% for GPT-5.4.

---

Use Cases

GPT-5.5 is particularly suitable for:

• Agentic Coding: Long-running autonomous engineering tasks, multi-file refactors, debugging across large codebases
• Knowledge Work: Spreadsheet modeling, operational research, document and slide generation
• Computer Use: Operating software, navigating UIs, and coordinating across tools
• Scientific Research: Bioinformatics, quantitative biology, mathematics, and multi-stage data analysis
• Financial and Legal Analysis: Report summarization, risk scoring, contract review
• Long Context Tasks: Document analysis and codebase reasoning with a 1M token input window

---

Related Links

• OpenAI Models Documentation
• Pricing Information
• Reasoning Guide
• Function Calling Guide
• Responses API Reference