GPT 5.6 Sol
GPT 5.6 Sol is the flagship of the GPT-5.6 family and the most capable of the three GPT-5.6 models on AI Gateway, with stronger agentic work across coding, biology, and cybersecurity than the previous generation, better token efficiency, and a context window of 1.1M tokens. Your use is subject to OpenAI's Terms & Privacy Policies.
import { streamText } from 'ai'
const result = streamText({ model: 'openai/gpt-5.6-sol', prompt: 'Why is the sky blue?'})Playground
Try out GPT 5.6 Sol by OpenAI. Usage is billed to your team at API rates. Free users (those who haven't made a payment) get $5 of credits every 30 days.
GPT 5.6 Sol
GPT 5.6 Sol
Providers
Route requests across multiple providers. Copy a provider slug to set your preference. Visit the docs for more info. Using a provider means you agree to their terms, listed under Legal.
| Provider |
|---|
| 1.1M | 128K | 2.7s | 45tps | $5/M+4 more | $30/M+4 more | Read:$0.5/M+2 more Write:$6.25/M+2 more | $10/K + input costs | +4 | 07/09/2026 | |||||
| 1.1M | 128K | 3.4s | 95tps | $5/M+1 more | $30/M+1 more | Read: $0.5/M+1 more Write: $6.25/M+1 more | $14/K + input costs | +3 | 07/09/2026 | |||||
| 272K | 272K | $5.50/M+1 more | $33/M+1 more | Read: $0.55/M+1 more Write: $6.88/M+1 more | $10/K + input costs | +3 | 07/09/2026 |
P50 throughput on live AI Gateway traffic, in tokens per second (TPS). Visit the docs for more info.
P50 time to first token (TTFT) on live AI Gateway traffic, in milliseconds. View the docs for more info.
Direct request success rate on AI Gateway and per-provider. Visit the docs for more info.
More models by OpenAI
| Model |
|---|
| 1.1M | 1.6s | 145tps | $1/M$0.20/M +1 more | $6/M$1.20/M +1 more | Read: $0.1/M$0.02/M+1 more Write: $1.25/M$0.25/M+1 more | $10/K + input costs | +4 |   | 07/09/2026 | ||||
| 400K | 0.5s | 138tps | $0.20/M | $1.25/M | Read:$0.02/M Write:— | $10/K + input costs | +4 | | 03/17/2026 | ||||
| 400K | 1.1s | 178tps | $0.75/M | $4.50/M | Read:$0.07/M Write:— | $10/K + input costs | +4 | | 03/17/2026 | ||||
| 400K | 5.4s | 147tps | $0.05/M | $0.40/M | Read:$0.01/M Write:— | $14/K + input costs | +3 | | 08/07/2025 | ||||
| 400K | 2.8s | 163tps | $0.25/M | $2/M | Read:$0.03/M Write:— | $14/K + input costs | +3 | | 08/07/2025 | ||||
| 131K | 0.2s | 481tps | $0.35/M | $0.75/M | Read:$0.25/M Write:— | — |   +5 | 08/05/2025 |
About GPT 5.6 Sol
GPT 5.6 Sol became available on AI Gateway on July 9, 2026, alongside GPT-5.6 Terra and GPT-5.6 Luna. All three GPT-5.6 models handle agentic work across coding, biology, and cybersecurity better than the previous generation, and all three use fewer tokens to reach an answer. GPT 5.6 Sol is the most capable of the set.
OpenAI positions GPT 5.6 Sol as the frontier model of the family for complex professional work. It corresponds to the unsuffixed tier of earlier GPT-5 families, so teams running a standard GPT-5.5 model today are looking at the direct successor. OpenAI routes the gpt-5.6 alias to GPT 5.6 Sol.
GPT 5.6 Sol accepts text and image input and returns text. The context window is 1.1M tokens, with up to 272K tokens of output per request. Reasoning tokens, streaming, function calling, and structured outputs are supported, along with web search, file search, code interpreter, computer use, and MCP (Model Context Protocol) tools through the Responses API. A fast mode is available when you want quicker responses from the same model, and it got faster after launch. See the live metrics on this page for current throughput.
Set the model to openai/gpt-5.6-sol in the AI SDK, the Chat Completions API, the Responses API, or another supported API format, from TypeScript or Python. AI Gateway routing rules can also rewrite an existing model slug to openai/gpt-5.6-sol, which moves production traffic onto GPT 5.6 Sol without a code change. List pricing is $5 per million input tokens and $30 per million output tokens, with cached input at $0.5. AI Gateway mirrors provider pricing with no markup and adds no platform fee on inference, including for Bring Your Own Key requests.
What To Consider When Choosing a Provider
- Configuration: GPT 5.6 Sol sits at the top of the GPT-5.6 lineup, so it carries the highest list price of the three. Route your hardest agentic and research requests here and send routine turns to GPT-5.6 Terra or GPT-5.6 Luna. A small routing layer keyed on task complexity keeps the cost profile predictable.
- Configuration: Token efficiency improved across the GPT-5.6 family, so a request can finish in fewer tokens than the previous generation needed. Pricing splits into short-context and long-context rates, and very large prompts bill at the long-context rate. Read the pricing panel on this page before you routinely fill the context window of 1.1M tokens.
- Zero Data Retention: AI Gateway supports Zero Data Retention for this model via direct gateway requests (BYOK is not included). To configure this, check the documentation.
- Authentication: AI Gateway authenticates requests using an API key or OIDC token. You do not need to manage provider credentials directly.
When to Use GPT 5.6 Sol
Best for
- Agentic Coding Workflows: Multi-step engineering tasks where smaller tiers stall and the flagship finishes the job
- Security and Biology Research: Two domains the GPT-5.6 family improved over the previous generation
- Long-Context Analysis: Whole repositories, filings, or research corpora that fit inside the context window
- Tool-Heavy Agents: Web search, file search, code interpreter, computer use, and MCP tools in one pipeline
- Quality-Critical Traffic: Requests where the answer ceiling decides the outcome rather than the unit cost
Consider alternatives when
- Everyday Production Traffic: GPT-5.6 Terra balances intelligence against cost across routine work
- High-Volume Cost Control: GPT-5.6 Luna carries the lowest price in the GPT-5.6 family
- Autonomous Coding Agents: GPT-5.x codex variants target sandboxed software engineering directly
- Live Voice Interfaces: The gpt-realtime family handles speech-to-speech conversation natively
Conclusion
GPT 5.6 Sol is the top of the GPT-5.6 lineup: the most capable of the three models, stronger on agentic coding, biology, and cybersecurity than the previous generation, and more token efficient. Send it the work a mid-tier model cannot close, and route the rest to GPT-5.6 Terra or GPT-5.6 Luna.