Loading page
OpenAI
GPT-5.5
GPT-5.5 is a chat-oriented large language model from OpenAI, offered through API providers tracked in the pricing table on this page.
Input from
$2.50 / 1M tokens
across 4 providers
API Pricing
Cheapest on OpenAI — 40% below avg| Provider | Input / 1M | Output / 1M | Cached / 1M |
|---|---|---|---|
| $2.50 | $15.00 | $0.250 | |
| $2.50 | $15.00 | $0.250 | |
| $5.00 | $30.00 | $0.500 | |
| $5.00 | $30.00 | $0.500 | |
| $5.00 | $30.00 | $0.500 | |
| $5.00 | $30.00 | $0.500 |
Prices updated daily. Last check: Sep 5, 2026
GPT-5.5 pricing by provider
Input, output, and batch rates, plus alternatives, for one provider at a time.
Performance & Benchmarks
Source: Artificial Analysis →Intelligence
45.6 / 100
Coding
74.9 / 100
Reasoning & Knowledge
- GPQA Diamond93.5%
- Humanity's Last Exam45.8%
Agentic & Tool Use
- Terminal-Bench Hard60.6%
- Terminal-Bench v2.184.3%
- τ²-bench93.9%
- τ-bench Banking39.0%
Instruction & Long Context
- IFBench75.9%
- Long-Context Reasoning84.3%
Benchmarks measured Sep 2026. Scores are independent evaluations, not vendor-reported.
Model Details
General
- Creator
- OpenAI
- Modalities
- Text
Capabilities
- Tool Calling
- No
- Open Source
- No
Strengths & Limitations
Strengths
- Part of OpenAI's GPT series, so it fits existing OpenAI-compatible client libraries and tooling with minimal integration work
- Chat-completions style interface with system prompts and multi-turn message history
- Suited to general-purpose text workloads rather than a single narrow task
- Available through providers listed in the pricing table, allowing side-by-side cost comparison
- OpenAI-compatible request format is widely supported by orchestration frameworks and agent libraries
- Naming places it within a well-documented family, making it straightforward to compare against sibling GPT models in this catalogue
Limitations
- We do not currently track a confirmed context window length for GPT-5.5, so maximum input size must be verified in OpenAI's documentation
- Our benchmark source has not reported output tokens per second or time-to-first-token for this model, so speed comparisons are unavailable from our data
- No verified benchmark scores in our database for reasoning, coding, or multilingual evaluation suites
- Supported input modalities are not confirmed in our metadata, so vision or audio capability should be checked before relying on it
- Provider availability may be narrower than for longer-established GPT models
Key Features
•Chat completions API interface with system, user, and assistant roles
•Multi-turn conversation handling with message history
•Member of OpenAI's GPT model series
•OpenAI-compatible request and response schema
•Streaming-style response consumption as supported by the serving provider
•Available through providers tracked on this page for price comparison
•General-purpose text generation, summarisation, and rewriting
About GPT-5.5
GPT-5.5 is a conversational large language model developed by OpenAI and part of the company's GPT series of general-purpose text models. It is catalogued here as a chat model, meaning it is designed to be used through a messages-style API where a system prompt and alternating user/assistant turns drive the interaction, rather than as a specialised embedding, image, or audio model.
Our database currently holds limited verified specifications for GPT-5.5. We do not track a confirmed context window length, modality list, or benchmark scores for this entry, so this page does not state figures for those attributes. The throughput and latency values reported by our benchmark source are not yet populated for this model, which means output speed and time-to-first-token comparisons against other GPT-series entries are not available from our data. Readers who need exact limits — maximum context, supported input types, tool-calling behaviour, or knowledge cutoff — should consult OpenAI's official model documentation for GPT-5.5.
In practice, models in this position within the GPT lineup are used for the same broad range of text workloads as their siblings: assistant-style chat, drafting and editing, summarisation, code assistance, and retrieval-augmented question answering. Because provider-level access terms and serving configurations differ, the practical trade-offs between GPT-5.5 and neighbouring GPT models often come down to which provider you use and what that provider charges — see the pricing table on this page for the current options.
Common Use Cases
GPT-5.5 is catalogued as a chat model, which makes it a candidate for assistant and copilot workloads: customer-facing conversational agents, internal knowledge assistants layered on top of a retrieval system, drafting and editing of documents and emails, summarising long-form text, and code explanation or generation inside developer tools. Because it uses OpenAI's chat-completions interface, it can usually be swapped into an existing pipeline that already targets a GPT model without rewriting prompt plumbing, which makes it convenient for A/B testing quality and latency against other entries in the family. For workloads with hard requirements — very long documents, image input, or strict latency budgets — confirm the model's context window, modalities, and serving performance in OpenAI's documentation first, since our metadata for GPT-5.5 does not yet include those values.
Frequently Asked Questions
How much does GPT-5.5 cost to use?
Pricing varies by provider and by pricing type — input tokens, output tokens, cached input, and batch or committed-capacity rates are often billed differently. Because rates change frequently, we do not quote figures in this write-up. See the pricing table on this page for the current per-provider rates we track for GPT-5.5.
What is GPT-5.5 best used for?
As a chat model, it suits assistant-style applications: multi-turn conversational agents, retrieval-augmented question answering, drafting and editing text, summarisation, and coding assistance. It is not an embedding, reranking, image, or speech model, so those tasks need a purpose-built model instead.
What context window does GPT-5.5 support?
Our database does not currently hold a verified context window figure for GPT-5.5, so we do not publish one here. Check OpenAI's official model documentation for the confirmed maximum context and output token limits before designing around a specific size.
Does GPT-5.5 accept image input?
We do not track confirmed modality information for this entry, so we cannot say either way. If your application depends on vision input, verify support in OpenAI's model documentation or with the specific provider you plan to use.
How does GPT-5.5 compare with other GPT models in this catalogue?
Meaningful comparison depends on benchmark and specification data, and our benchmark source has not yet reported throughput or latency numbers for GPT-5.5. For now the most reliable comparison you can make from this site is on cost and provider availability, using the pricing table above alongside the pages for other GPT-series models.