GPT-5.5

gpt-5.5

GPT-5.5 is a frontier model designed by OpenAI for complex professional workloads, building on GPT-5.4 with stronger reasoning capabilities, higher reliability, and improved token efficiency on difficult tasks. It features a 1M+ token context window (922K input, 128K output), supports text and image inputs, and enables large-scale reasoning, coding, and multimodal workflows within a single system.

Context window1.1M

ProviderOpenAI

Released2026/04/25

Playground

Pricing

The same model is available through multiple service channels — choose based on latency, reliability and cost.

Prices in $ / 1M tokens

To pick a channel, add a provider field to the request body, for example "provider": { "channel": "direct" }. Valid values are direct / stable / economical; omit it to use the default channel.

Direct

Direct upstream connection — best when you need native behavior and the full context window.

Input context	Input	Output	Cache read
≤ 272K	5.00/M	30.00/M	0.50/M
> 272K	10.00/M	45.00/M	1.00/M

Overview

Input

Text Image

Output

Text

GPT-5.5 API: A Frontier Model for Complex Professional Workloads

GPT-5.5 is a frontier model released by OpenAI in April 2026, building on GPT-5.4 with enhanced reasoning, reliability, and token efficiency on difficult tasks. It features a context window of over 1 million tokens (approximately 920K input + 128K output), supports text and image inputs, and can handle large-scale reasoning, coding, and multimodal workflows within a single system.

In public benchmarks, GPT-5.5 scored 82.7% on Terminal-Bench 2.0, 84.9% on GDPval, 58.6% on SWE-Bench Pro, and 51.7% on FrontierMath Tier 1–3. The main improvement direction for this generation is "doing the same task correctly with fewer tokens."

SeaWhale AI provides GPT-5.5 through an OpenAI-compatible interface, with support for tool calling, reasoning effort control, streaming output, and multimodal input.

Get API Key · Model ID: gpt-5.5

Why Choose GPT-5.5

Mature and stable frontier tier — Released earlier than GPT-5.6, with a more mature ecosystem and richer prompt-engineering experience
1M+ token context — Approximately 920K input tokens, covering full codebases and long document collections
Greater token efficiency — Compared with GPT-5.4, achieves equal or better results on difficult tasks with fewer tokens
Strong terminal and coding performance — 82.7% on Terminal-Bench 2.0, reliable on long-horizon coding tasks
Multimodal input — Unified processing of text and images

Core Capabilities

01 Complex Reasoning and Long-Horizon Coding

GPT-5.5 is designed for "high-risk, multi-step" work: cross-codebase refactoring, multi-stage debugging, and engineering tasks that require precise execution. Its 82.7% on Terminal-Bench 2.0 demonstrates its execution capability in real terminal environments.

Planning, implementation, and verification across codebases
Multi-stage debugging and root-cause localization
Long-horizon agentic execution

02 1M+ Token Context

A window of approximately 920K input tokens means you can fit an entire medium-to-large codebase, a complete research corpus, or hundreds of pages of documents at once, without complex chunking or retrieval.

Full codebase analysis
Cross-document reasoning over long document collections
Up to 128K output tokens

03 Professional Knowledge Work

The 84.9% on GDPval reflects the model's performance on real-world professional tasks—report writing, financial analysis, research synthesis, and structured deliverable generation.

Research reports and financial analysis
Structured document and table output
Synthesis and distillation of long source materials

04 Tool Calling and Multimodality

Supports function calling, parallel tool execution, streaming output, and image input, and can be directly integrated into existing agent frameworks.

Function calling and tool orchestration
Streaming output and incremental parameters
Image understanding and chart analysis

Best Use Cases

Scenario	Description
Codebase-level engineering	Large-scale refactoring, cross-file feature implementation and verification
Long-horizon agents	Multi-step planning, tool calling, and result verification
Long-context analysis	Cross-document reasoning within a 920K-token input window
Professional knowledge work	Financial analysis, research synthesis, report writing
Mathematical and scientific reasoning	High-difficulty reasoning tasks such as FrontierMath
Multimodal workflows	Mixed text-image processing of charts, screenshots, and documents

Differences Between GPT-5.5, GPT-5.6 Sol, and GPT-5.5 Pro

Capability	GPT-5.5	GPT-5.5 Pro	GPT-5.6 Sol
Model ID	`gpt-5.5`	`gpt-5.5-pro`	`gpt-5.6-sol`
Positioning	Frontier general-purpose tier	Deliberate research tier	Next-generation flagship
Context window	1M+ tokens	1M+ tokens	100K tokens
Max output	128K tokens	128K tokens	256K tokens
Terminal-Bench	82.7% (2.0)	Higher	Top tier (2.1)
Best for	General frontier workloads	Long-horizon research and full-repo refactoring	Complex agents and frontier coding

Specific billing is subject to the real-time price card at the top of the page.

How to Use the GPT-5.5 API

1. Create a SeaWhale AI API key Generate a key in the console and top up your balance.

2. Decide between 5.5 and 5.6 Choose GPT-5.5 if you need an ultra-long input context (920K tokens); choose GPT-5.6 Sol if you need stronger coding-agent performance and longer output.

3. Call the API

curl -X POST https://api.seawhaleai.com/v1/chat/completions \
  -H 'Authorization: Bearer YOUR_API_KEY' \
  -H 'Content-Type: application/json' \
  -d '{
    "model": "gpt-5.5",
    "messages": [
      {"role": "user", "content": "Read the entire repository, identify performance bottlenecks, and provide optimization plans and verification methods."}
    ],
    "stream": true
  }'

FAQ

When was GPT-5.5 released? Released on April 23, 2026, with API access opening the following day.

How large is the context window exactly? Over 1 million tokens, with approximately 920K available for input and 128K for output—one of the largest input windows in the current GPT series.

GPT-5.5 or GPT-5.6? If your bottleneck is input length, GPT-5.5's 920K-token input window is more advantageous; if the bottleneck is coding-agent capability or output length, the GPT-5.6 series is more suitable.

How do the key benchmarks look? Terminal-Bench 2.0 82.7%, GDPval 84.9%, SWE-Bench Pro 58.6%, FrontierMath Tier 1–3 51.7%, Tier 4 35.4%.

What input modalities are supported? Text and image. Tool calling, streaming output, and reasoning effort control are all available.

What's the difference from GPT-5.5 Pro? Pro is a deliberate variant aimed at long-horizon research and cross-repo refactoring, with deeper reasoning but significantly higher cost. For regular work, the standard version is sufficient.

Why Choose SeaWhale AI for the GPT-5.5 API

Ultra-long input window, ready to use — No need to build your own chunking and retrieval pipeline
OpenAI-compatible interface — Integrate with existing SDK code by changing just two lines
Multiple model generations coexist — GPT-5.5 and GPT-5.6 series share the same API key, allowing easy comparison and switching
Unified billing — Centralized management of usage and costs across models

API

API integration

Model IDUse this value as the model in inference requests

gpt-5.5

API KeyBearer token used to authenticate inference requests

Manage API keys

Base URLOpenAI compatible · /chat/completions

https://api.seawhaleai.com/v2/

provider OptionalSelects a service channel; omit it and the system picks the default

"provider": { "channel": "direct" }

gpt-5.5 usage examples

SeaWhale AI is compatible with the OpenAI API protocol, so you can call it with the OpenAI SDK or plain HTTP requests. Streaming is enabled by default.

About the provider parameter (optional, a SeaWhale AI extension): most models are served over several channels that differ slightly in price and reliability. Add a provider field to the request body to pick one; omit it and the system selects the default channel — normal calls are unaffected.

provider is not part of the official OpenAI protocol — it is a SeaWhale AI extension that only takes effect on this platform. The OpenAI SDK allows custom fields like this to pass through; see the examples below.

Value	Channel	Best for
`direct`	Direct	The official upstream link, for native behavior and the full context window
`stable`	Preferred	Balanced availability and speed — a good default for production traffic
`economical`	Economy	Cost first, well suited to batch processing and price-sensitive workloads

Available channels and their prices are listed under "Pricing" above (channels vary by model). Additional notes:

Syntax: "provider": { "channel": "direct" }.
If the requested channel is not enabled for that model, the request automatically falls back to the default channel and returns normally — it does not error.
With the official SDKs: in Python pass it via extra_body; in Node.js put it directly on the request object and it passes through. In TypeScript projects, add a // @ts-expect-error line to skip the type check.

cURLPythonNode.js

curl https://api.seawhaleai.com/v2/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer <API_KEY>" \
  -d '{
    "model": "gpt-5.5",
    "messages": [
      {"role": "system", "content": "You are a helpful assistant."},
      {"role": "user", "content": "Hello!"}
    ],
    "provider": { "channel": "direct" },
    "stream": true
  }'
# provider is optional — remove this line to use the default channel

from openai import OpenAI

client = OpenAI(
    base_url="https://api.seawhaleai.com/v2",
    api_key="<API_KEY>",
)

stream = client.chat.completions.create(
    model="gpt-5.5",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Hello!"},
    ],
    stream=True,
    # Optional: pick a service channel; omit to use the default
    extra_body={"provider": {"channel": "direct"}},
)

for chunk in stream:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)

import OpenAI from 'openai'

const client = new OpenAI({
  baseURL: 'https://api.seawhaleai.com/v2',
  apiKey: '<API_KEY>',
})

const stream = await client.chat.completions.create({
  model: 'gpt-5.5',
  messages: [
    { role: 'system', content: 'You are a helpful assistant.' },
    { role: 'user', content: 'Hello!' },
  ],
  stream: true,
  // Optional: pick a service channel; omit to use the default
  // @ts-expect-error provider is a SeaWhale AI extension, not in the OpenAI SDK types
  provider: { channel: 'direct' },
})

for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content ?? '')
}