GPT-5.5 pricing and specs
GPT-5.5 pricing and specsDefault frontier GPT for coding, computer use, research, and knowledge work
GPT-5.5 price per 1M tokens
| Type | ₽ / 1M | $ / 1M |
|---|---|---|
| Input tokens | 423 ₽ | $5 |
| Output tokens | 2 535 ₽ | $30 |
| Cache read | 42,25 ₽ | $0.5 |
Prices in rubles as of 18.09.2026 at the CBR rate: 1 $ = 84.51 ₽.
GPT-5.5 specs and limits
- Context window
- 1.1M
- Max output
- 128K
- Max input
- 922K
- Input
- text, image, pdf
- Output
- text
- Family
- gpt
- Released
- 2026-04-23
- Updated
- 2026-04-23
- Knowledge cutoff
- 2025-12-01
- Also known as
- chatgpt .5, chat gpt .5, openai .5, gpt .5, gpt-5 .5, gpt-5.6 .5
What GPT-5.5 can do
- ✓Reasoning
- ✓Tool calling
- ✓Structured output
- ✓Attachments / files
- ✓Vision (images)
- —Temperature control
- —Open weights
GPT-5.5 benchmarks
| Benchmark | Score | Metric |
|---|---|---|
| SWE-Bench Pro | 58.6 | resolve rate |
| Terminal-Bench | 78.2 | success rate |
| SWE-Atlas Codebase QnA | 45.43 | score |
| SWE-Atlas Refactoring | 44.79 | score |
| SWE-Atlas Test Writing | 42.59 | score |
| Artificial Analysis Coding Agent Index | 65.3 | average pass@1 |
| SWE-Atlas Codebase QnA | 80.8 | pass@1 |
| SWE-Bench Pro | 30.9 | pass@1 |
| Terminal-Bench | 84.1 | pass@1 |
| Artificial Analysis Coding Agent Index | 60.4 | average pass@1 |
| SWE-Atlas Codebase QnA | 79.1 | pass@1 |
| SWE-Bench Pro | 26.2 | pass@1 |
| Terminal-Bench | 75.8 | pass@1 |
| Artificial Analysis Coding Agent Index | 57.8 | average pass@1 |
| SWE-Atlas Codebase QnA | 75 | pass@1 |
| SWE-Bench Pro | 24.9 | pass@1 |
| Terminal-Bench | 73.4 | pass@1 |
| Terminal-Bench | 82.7 | success rate |
| GPQA Diamond | 93.6 | accuracy |
| Humanity's Last Exam | 41.4 | accuracy |
| Humanity's Last Exam | 52.2 | accuracy |
| OSWorld-Verified | 78.7 | success rate |
| BrowseComp | 84.4 | accuracy |
| MMMU Pro | 81.2 | accuracy |
| ARC-AGI-2 | 85 | accuracy |
| FrontierMath | 51.7 | accuracy |
| FrontierMath | 35.4 | accuracy |
| GDPval | 84.9 | wins or ties |
| MCP Atlas | 75.3 | success rate |
| Toolathlon | 55.6 | success rate |
| τ²-Bench Telecom | 98 | success rate |
Frequently asked questions
How much does the GPT-5.5 API cost?
Input costs $5 per 1M tokens (423 ₽) and output costs $30 per 1M (2,535 ₽). OpenAI bills the prompt and the completion separately, so the price of a call depends on how long both are. Ruble figures use the Bank of Russia rate of 84.51 ₽ per dollar for September 2026.
How much do 1,000 GPT-5.5 tokens cost?
List prices are quoted per 1M tokens, so 1,000 tokens cost a thousandth of that: divide $5 for input and $30 for output by 1,000. The same rates in rubles are 423 ₽ and 2,535 ₽ per 1M tokens. Both prompt tokens and completion tokens are billed.
How many tokens does GPT-5.5 hold?
The GPT-5.5 context window is 1.1M tokens (1,050,000). Everything shares that budget: the system prompt, the dialogue history, attached files and the answer the model writes. The figure comes from the OpenAI model card as of September 2026.
What are the GPT-5.5 limits?
The context window is 1.1M tokens (1,050,000) and a single response is capped at 128K tokens, so longer output has to be generated in parts. Rate limits are set by OpenAI per account and depend on your plan rather than on the model, so they are not listed here.
How is the GPT-5.5 price in rubles calculated?
Providers quote their tariffs in US dollars, so the ruble amounts here are a conversion at the official Bank of Russia rate of 84.51 ₽ per dollar. For GPT-5.5 that gives 423 ₽ per 1M input tokens and 2,535 ₽ per 1M output tokens. When the rate moves, the ruble price moves with it.
How is GPT-5.5 different from other ChatGPT models?
GPT-5.5 belongs to the GPT family at OpenAI. Models in the ChatGPT line differ in context size, price per 1M tokens and supported capabilities, so compare them by the numbers on their cards rather than by the name. Every ChatGPT model is listed on the provider page.
Can GPT-5.5 reason?
Yes, GPT-5.5 declares a reasoning mode: it works through intermediate steps before it answers. Those steps add output tokens, so a call with a long reasoning chain costs more and takes longer than a plain completion. Capabilities declared for the model: reasoning, tool calling, image input, structured output.
What capabilities does GPT-5.5 support?
The declared capabilities of GPT-5.5 are: reasoning, tool calling, image input, structured output. It accepts text, images, PDF as input, and every attachment consumes tokens from the shared context window of 1.1M tokens. The list comes from the OpenAI model card and is refreshed with the catalog as of September 2026.