# OpenAI: gpt-5.6-sol

- Model ID: `gpt-5.6-sol`
- Provider: OpenAI
- Web version: https://www.moleapi.com/en/models/openai/gpt-5.6-sol
- Content status: deep

## Model introduction

GPT-5.6 Sol is OpenAI's flagship GPT-5.6 model for complex professional work, with an emphasis on difficult reasoning, software engineering, research, and tool-heavy agent workflows.

## Model capabilities

- OpenAI compatible
- Responses API
- Anthropic compatible
- Gemini compatible
- Image generation
- Vision
- Prompt cache
- Reasoning

### Verified specifications

- Official positioning: GPT-5.6 flagship for complex professional work
- Context window: 1,050,000 tokens
- Maximum output: 128,000 tokens
- Knowledge cutoff: 2026-02-16
- Release date: 2026-07-09
- Input / output: Text, image / Text
- Reasoning levels: none, low, medium, high, xhigh, max, and Pro
- GPT-5.6 additions: Explicit caching, persisted reasoning, programmatic tools, multi-agent beta
- Alias: gpt-5.6 routes to gpt-5.6-sol
- Verified capabilities: Functions, structured outputs, streaming, web and file search, code execution, computer use, MCP

### Independent capability evaluation

- Reasoning, xhigh effort
- Intelligence: 58 (#6/190)
- Output speed: 60.7 tok/s (#93/190)
- Time to first token: 69.2s
- Source: https://artificialanalysis.ai/models/gpt-5-6-sol-xhigh

The xhigh profile places near the frontier on composite capability, with a long measured first-token delay that makes effort selection important in production.

## Model pricing and access

Pricing is supplied dynamically by the MoleAPI console API.

### Standard (default, x1)

| Tier | Condition | Input | Output | Cache read | Cache write | Image input | Image output |
| --- | --- | ---: | ---: | ---: | ---: | ---: | ---: |
| standard | Length ≤ 272K | $5 | $30 | $0.5 | $6.25 | $8 | $30 |
| long_context | - | $10 | $45 | $1 | $12.5 | $8 | $30 |

### Discount (discount, x0.8)

| Tier | Condition | Input | Output | Cache read | Cache write | Image input | Image output |
| --- | --- | ---: | ---: | ---: | ---: | ---: | ---: |
| standard | Length ≤ 272K | $4 | $24 | $0.4 | $5 | $6.4 | $24 |
| long_context | - | $8 | $36 | $0.8 | $10 | $6.4 | $24 |

### Relay (relay, x0.3)

| Tier | Condition | Input | Output | Cache read | Cache write | Image input | Image output |
| --- | --- | ---: | ---: | ---: | ---: | ---: | ---: |
| standard | Length ≤ 272K | $1.5 | $9 | $0.15 | $1.875 | $2.4 | $9 |
| long_context | - | $3 | $13.5 | $0.3 | $3.75 | $2.4 | $9 |

### Temporary (availability may vary) (temp, x0.1)

| Tier | Condition | Input | Output | Cache read | Cache write | Image input | Image output |
| --- | --- | ---: | ---: | ---: | ---: | ---: | ---: |
| standard | Length ≤ 272K | $0.5 | $3 | $0.05 | $0.625 | $0.8 | $3 |
| long_context | - | $1 | $4.5 | $0.1 | $1.25 | $0.8 | $3 |

### Access protocols

| Protocol | Method | Endpoint |
| --- | --- | --- |
| openai | POST | /v1/chat/completions |
| openai-response | POST | /v1/responses |
| anthropic | POST | /v1/messages |
| gemini | POST | /v1beta/models/{model}:generateContent |
| image-generation | POST | /v1/images/generations |
| openai-response-compact | POST | /v1/responses/compact |
| openai-alpha-search | POST | /v1/alpha/search |

- Live pricing source: https://home.moleapi.com/api/pricing

## About gpt-5.6-sol

GPT-5.6 Sol is OpenAI’s flagship GPT-5.6 model for complex production work, including coding, agentic tool use, computer use, and long-horizon knowledge workflows. [3] It combines configurable reasoning with Responses API features for tool orchestration and multi-turn applications. [4]

## Positioning and core capabilities

GPT-5.6 Sol is the frontier-capability tier in the GPT-5.6 family, while Terra is positioned for balanced everyday work and Luna for efficient high-volume workloads. [3] The GPT-5.6 family is generally available. [3]

Sol supports six reasoning-effort levels: `none`, `low`, `medium`, `high`, `xhigh`, and `max`. [4] Pro mode is a separate reasoning setting that prioritizes reliability over latency and token use, with `medium` as the default effort when no level is specified. [4]

OpenAI recommends the Responses API for GPT-5.6 reasoning, tools, and multi-turn workflows. [4] GPT-5.6 adds explicit prompt caching and persisted reasoning as separate API features. [4]

## Capabilities and specs

GPT-5.6 Sol accepts text and image inputs and produces text outputs. [1] Its tool-oriented API features include Programmatic Tool Calling and beta multi-agent orchestration. [4]

| Specification | Details |
| --- | --- |
| Input modalities | Text and image [1] |
| Output modality | Text [1] |
| Reasoning efforts | `none`, `low`, `medium`, `high`, `xhigh`, and `max` [4] |
| Reasoning mode | Pro mode is available as a reasoning configuration. [4] |
| Workflow features | Explicit prompt caching, persisted reasoning, Programmatic Tool Calling, and beta multi-agent orchestration. [4] |

Programmatic Tool Calling enables the model to write JavaScript that invokes eligible tools and processes intermediate results in a hosted runtime. [4] The Responses API multi-agent beta enables a GPT-5.6 instance to coordinate parallel subagents and synthesize their results. [3]

OpenAI reports vendor-published GPT-5.6 Sol results of 92.2% on BrowseComp and 62.6% on OSWorld 2.0. [3]

## Use cases

OpenAI positions GPT-5.6 Sol for difficult professional work spanning software engineering, research, computer use, cybersecurity, science, design judgment, and workflows that coordinate several tools. [3]

- Complex coding and software-engineering workflows, including work that benefits from extended reasoning. [3]
- Long-horizon research and professional knowledge tasks. [3]
- Tool-heavy agent workflows using Programmatic Tool Calling and persisted reasoning. [4]
- Parallel workstreams coordinated through the Responses API multi-agent beta. [3]
- Frontend and interface work involving layout, visual hierarchy, and design judgment. [4]
- Authorized cybersecurity work such as review, patching, and threat modeling. [3]

## Material limitations

GPT-5.6 uses real-time cybersecurity and biology misuse classifiers that can refuse, block, or pause some requests. [3] These controls can occasionally affect legitimate dual-use cybersecurity or biology work. [3]

Higher reasoning efforts and Pro mode prioritize result quality over latency and token use. [4] OpenAI describes `max` as extending reasoning beyond `xhigh`, while multi-agent coordination increases token use when parallel workstreams are used. [3][4]

OpenAI states that GPT-5.6 models do not cross its Critical capability threshold in cybersecurity or biology. [3]

## Sources

- [1] [OpenAI GPT-5.6 Sol model documentation](https://developers.openai.com/api/docs/models/gpt-5.6-sol)
- [3] [OpenAI GPT-5.6 announcement](https://openai.com/index/gpt-5-6/)
- [4] [OpenAI GPT-5.6 usage and migration guide](https://developers.openai.com/api/docs/guides/latest-model)


## Code examples

### cURL

```bash
curl https://api.moleapi.com/v1/responses \
  -H "Authorization: Bearer $MOLEAPI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-5.6-sol","input":"Explain this problem step by step."}'
```

### Python

```python
from openai import OpenAI

client = OpenAI(base_url="https://api.moleapi.com/v1", api_key="YOUR_MOLEAPI_API_KEY")
response = client.responses.create(model="gpt-5.6-sol", input="Explain this problem step by step.")
print(response.output_text)
```

### TypeScript

```typescript
import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://api.moleapi.com/v1", apiKey: process.env.MOLEAPI_API_KEY });
const response = await client.responses.create({ model: "gpt-5.6-sol", input: "Explain this problem step by step." });
console.log(response.output_text);
```

## More related models

- [gpt-5.6-luna](https://www.moleapi.com/en/models/openai/gpt-5.6-luna) — OpenAI
- [gpt-5.6-terra](https://www.moleapi.com/en/models/openai/gpt-5.6-terra) — OpenAI
- [gpt-5.5](https://www.moleapi.com/en/models/openai/gpt-5.5) — OpenAI

## Frequently asked questions

### How is gpt-5.6-sol priced?

Prices are read from the MoleAPI console API and update with model prices, context tiers, and account groups.

### How can I access gpt-5.6-sol?

The catalog currently lists openai, openai-response, anthropic, gemini, image-generation, openai-response-compact, openai-alpha-search.

### How do I switch an existing project to gpt-5.6-sol?

Keep the MoleAPI API address and key, replace the model parameter, and check protocol-specific parameters.

### Where do the gpt-5.6-sol details come from?

Capabilities and limitations are checked against OpenAI and the other cited pages; pricing and protocols come from the MoleAPI console.