GPT-5.5
⧉GPT-5.5 提供业界领先的对话、推理与代码能力，支持流式输出、长上下文与结构化任务。完全兼容 OpenAI SDK，可直接接入 Codex / Cursor / Claude Code 等工具，零改造迁移。
Live Test · Playground
Try out GPT-5.5 right here (available after login)。
Input
Advanced
Conversation
About GPT-5.5
GPT-5.5 is OpenAI's flagship chat and reasoning model, built for multi-step reasoning, long-running agent workflows, and reliable structured output where answer quality matters most. On NezhaGate it is served through an OpenAI-compatible endpoint: point your base_url at NezhaGate and set model to "gpt-5.5" to reuse the OpenAI SDK you already have. Streaming is supported, and billing is pay-as-you-go, with failed requests not charged.
Use cases
Tackle multi-step derivations, logical arguments, and data insights, producing explainable, traceable conclusions on problems that require careful decomposition.
Acts as the reasoning core of an agent, planning steps, calling external tools, and combining results across long, multi-turn decision flows.
Reliably emits JSON and other structured results that plug straight into downstream systems, ideal for data extraction, form filling, and function calling.
Drafts reports, proposals, technical docs, and marketing copy while keeping logic coherent and tone consistent across long pieces.
Powers support bots, knowledge Q&A, and product assistants, tracking context across turns to interpret intent and reply with high quality.
How to choose
Reach for the flagship GPT-5.5 when tasks involve multi-step reasoning, complex agent orchestration, or demand the highest output quality; choose GPT-5.6 Terra when you want a balance of speed and cost, and GPT-5.6 Luna for lightweight, latency-sensitive workloads. Comparing flagships across providers, Claude Sonnet 4.6 leans toward long-form reasoning and code while Gemini 3.1 Pro leans toward Google-ecosystem reasoning—all available behind the same NezhaGate endpoint, so you can switch the model field to try each.
FAQ
What is GPT-5.5?
Does GPT-5.5 support streaming?
How do I call gpt-5.5 with the OpenAI SDK?
Which use cases fit GPT-5.5 best?
GPT-5.5 vs GPT-5.4: how do I choose?
Related Models
Explore other models you can integrate.
GPT-5.6 Sol
前沿旗舰 Agentic 编程模型
GPT-5.6 Terra
日常均衡 Agentic 编程模型
GPT-5.6 Luna
快而省的 Agentic 编程模型
GPT Image 2
高质量文生图 / 图生图模型
Nano Banana 2
高质量文生图 / 图生图模型
Nano Banana Pro
旗舰级文生图 / 图生图模型
Claude Sonnet 4.6
均衡高效的对话 / 代码模型
Gemini 3.1 Pro
Google 新一代旗舰推理模型
Gemini 3.6 Flash
新一代高速思考模型 · 四档思考预算
Gemini 3.6 Flash High
深度思考档
Gemini 3.6 Flash Low
极速低耗档
Gemini 3.6 Flash Tiered
自动调档
Gemini 3 Flash
高速低延迟，高性价比对话模型
Gemini 3.5 Flash
新一代高速思考模型
Gemini 2.5 Flash
高速低延迟，高性价比对话
Veo 3.1
文生 / 图生视频 · 多档 · 多分辨率
