GPT Image 2.5 is live — OpenAI's newest image model, targeted edits that leave the rest of the frame alone

GPT-5.6 Sol — Frontier Tier, 1.05M Context, Max Effort

GPT-5.6 Sol is the frontier model of the GPT-5.6 family and what the bare `gpt-5.6` alias resolves to. GPT-5.6 Sol reads a 1,050,000-token context window, returns up to 128,000 tokens, and carries the family's full six-rung reasoning ladder including the new max setting. OpenAI rates its reasoning Highest and its speed Fast. On reAPI, GPT-5.6 Sol bills below OpenAI's published rate on both input and output.

Highlights
Reasoning effort tiers · tools & JSON mode
Best for
Coding, hard reasoning, review
Input
Chat messages (OpenAI-compatible)
Output
Text, streaming optional
GPT-5.6 Sol

Ask GPT-5.6 Sol

Start with the prompt in the composer or choose an example.

Enter to send · Shift+Enter for a new line

Pricing

Credit-based — 1 credit = $0.001 USD. Pay only for completed generations.

First test cost
$0.019219.2 credits

One 1K-token prompt with a 1K-token answer at the cheapest tier.

Testing budget guide
Add credits
$10
≈ 520 tests
$50
≈ 2604 tests
$100
≈ 5208 tests
Category
Token pricing
Long context · prompts over 272K tokens
Input
$3.20
Price · 1M tokens
$6.40
Price · 1M tokens
Output
$16.00
Price · 1M tokens
$24.00
Price · 1M tokens
Cache read
$0.32
Price · 1M tokens
$0.64
Price · 1M tokens
Cache write
$4.00
Price · 1M tokens
$8.00
Price · 1M tokens

What is the GPT-5.6 Sol API?

GPT-5.6 Sol is the frontier tier of OpenAI's GPT-5.6 family and what the bare gpt-5.6 alias resolves to, served on reAPI through an OpenAI-compatible /api/v1/chat/completions endpoint. It reads a 1,050,000-token context window, returns up to 128,000 tokens, and carries the family's full six-rung reasoning ladder including the new max setting, with OpenAI rating its reasoning Highest and its speed Fast. Billing is per 1M tokens, with input and output priced separately; reasoning tokens count as output.

What you can build with this model

Real-world workflows and production use cases you can build and ship with this model.

GPT-5.6 Sol tracing a defect through a whole service before proposing a fix

The debugging and review work that has to be right

OpenAI's guidance for the top of the effort ladder names security and code review, complex debugging and deep planning — tasks where quality matters more than latency. GPT-5.6 Sol is the tier those settings were built for, and the shape of the work follows from the limits: a 1,050,000-token window means the service, its tests and its recent history all go into one request, so the model is reasoning about the actual system rather than a summary of it. Ask GPT-5.6 Sol to prove a cause before proposing a fix and the deep pass earns its cost; ask it to guess quickly and you are paying frontier rates for a cheap tier's job.

Read the API docs
GPT-5.6 Sol running a long multi-tool agentic task

Long-horizon agentic runs that do not drift

Deep research, asynchronous workflows and agentic tasks that require long runs are OpenAI's stated cases for xhigh, and this generation adds the pieces those runs need: programmatic tool calling, explicit prompt-caching controls and persisted reasoning. With functions, web search, file search and computer use all available, GPT-5.6 Sol is the tier you point at a task that will take many steps and many tool calls, where the failure you are buying protection against is not a wrong token but a plan that quietly falls apart at step thirty.

GPT-5.6 Sol weighing a high-stakes decision with full evidence in context

High-value judgement, where the cheap tier is the expensive choice

There is a class of request where a wrong answer costs more than the whole month's inference bill: a migration plan, a contract review, an architectural commitment, a security finding. On those, tier selection is not a cost decision. GPT-5.6 Sol is the model OpenAI puts at the top of this family — reasoning rated Highest, and the target of the bare `gpt-5.6` alias, which is a fair signal of where OpenAI expects the default to sit. Route the volume to a cheaper sibling and keep GPT-5.6 Sol for the calls you would want a senior engineer to make.

Why reAPI for GPT-5.6 Sol

Below the published rate

OpenAI publishes a per-million-token rate for GPT-5.6 Sol input and output. On reAPI, GPT-5.6 Sol bills below both, and the rates on this page are updated live. No subscription underneath it, no minimum commitment, pay-as-you-go from your reAPI balance. The exact numbers are in the pricing table on this page.

OpenAI-compatible drop-in

GPT-5.6 Sol is served from a standard `/api/v1/chat/completions` endpoint. If your code already speaks that shape, adopting GPT-5.6 Sol is a base URL, a key and a model string — the same SDKs work once the base URL points at reapi.ai. No SDK swap, no bespoke client.

One key across the whole ladder

The same reAPI key that calls GPT-5.6 Sol calls its cheaper siblings, the Claude and Gemini families, and the open-weight tiers. That is what makes a frontier tier affordable in practice: send the routine turns somewhere cheap and escalate to GPT-5.6 Sol by changing a model string, not by signing up with a second vendor.

GPT-5.6 Sol vs Claude Opus 5

Two frontier tiers, the same order of magnitude on price, and genuinely different contracts. GPT-5.6 Sol gives you a six-rung effort ladder that includes turning reasoning off entirely, image input, and a million-token window with a 128K output ceiling. Claude Opus 5 reasons by default, refuses the sampling parameters outright, and takes PDF rather than images. Neither is the general answer; the rows below are the ones that decide it for a given route.

Capability
GPT-5.6 Sol on reAPI
Claude Opus 5
Vendor positioning
Frontier model for complex professional work; reasoning rated Highest in its family
For complex agentic coding and enterprise work
Context and output limits
1,050,000-token context, 128,000-token maximum output
1M-token context, 128k maximum output
Reasoning default
Effort defaults to medium, and none turns reasoning off entirely
Thinking on by default; can be disabled only at effort high or lower
Effort ladder
Six rungs — none, low, medium, high, xhigh, max
Five rungs from low through max, defaulting to high
Sampling parameters
Reasoning effort is the intended control on this family
Sampling parameters removed — sending them returns an error
Non-text input
Images, at their original dimensions
Images and PDF

Comparison reflects behavior documented by each vendor at the time of writing. Rates are described as relationships rather than amounts; the live numbers for GPT-5.6 Sol are in the pricing table above.

Ship GPT-5.6 Sol in three steps

  1. step 01

    Create a reAPI API key

    Sign in to reAPI and create a key under API keys. That one key reaches GPT-5.6 Sol and every other model on the platform, billed from a single balance.

    Open
  2. step 02

    Point your client at reapi.ai

    Set your base URL to reapi.ai's API and your key to the one you just created. Any client that already speaks OpenAI Chat Completions works unchanged — no SDK swap is needed to reach GPT-5.6 Sol.

    Open
  3. step 03

    Send `gpt-5.6-sol` as the model

    Set the model field to `gpt-5.6-sol` — with the dots — and post your messages. Turn on streaming, since a deep pass can be slow to first token, and set reasoning effort deliberately: the family default is medium, and this tier is worth paying for at high and above.

    Open
docs/api/gpt-5-6-sol

API reference

Drop-in code and the full parameter table.

View full reference
curl https://reapi.ai/api/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.6-sol",
    "messages": [
      { "role": "user", "content": "Hello" }
    ],
    "reasoning_effort": "high",
    "stream": true
  }'

Parameter validation failed

Rejected synchronously with the failing field named. Check enum values, ranges, and URL fields — nothing is charged.

Authentication or balance issue

401 means a missing or invalid Bearer key; 402 means the reserve exceeds your balance. Manage keys and credits in the console.

Content or material rejected

Reference material and outputs pass automated moderation. Rejected tasks fail with a clear error and a full refund.

Task failed or timed out

A task that reaches a failed state is never charged — the reserve refunds automatically. Keep the task id and retry when ready.

Frequently asked questions

Common questions about this model.

GPT-5.6 Sol is billed pay-as-you-go per token from your reAPI balance, with separate input and output rates and no subscription. Both sit below OpenAI's published per-token rate. Current numbers are in the pricing table on this page.

Related models

Explore more models in the same category.

View all models
start building

Ready to ship?

Try it in the playground or grab an API key to integrate now.

Get API keyView API docs