# GPT-6 Sol - Core.Today AI API

> The mid tier of OpenAI's GPT-6 generation (in the API since September 14, 2026) — strong reasoning, coding and multimodal performance at $2/$10 per million tokens, a fifth of GPT-6 Astra. 1.05M token context window, 128K max output tokens, cached inputs at a 90% discount. Prompts above 272K input tokens are billed at 2x input / 1.5x output for the whole request.

- **Provider**: OpenAI
- **Model ID**: gpt-6-sol
- **Category**: LLM
- **Credits**: 2 per request
- **Speed**: Fast
- **Quality**: Ultra

## Model Specifications
- **Context Window**: 1.1M tokens
- **Max Output**: 128K tokens
- **Training Cutoff**: 2026-04
- **Supported Formats**: text, json, markdown
- **Compatible SDK**: OpenAI

### Capabilities
- Vision (image input)
- Function Calling
- Streaming
- JSON Mode
- System Prompt

### Token Pricing (per 1M tokens)
- **Input Tokens**: 3,716 credits ($2.48)
- **Output Tokens**: 18,580 credits ($12.39)
- **Cached Tokens**: 371.6 credits ($0.25)

## Features
- Mid tier of the GPT-6 generation
- 1.05M token context window
- 128K max output tokens
- Knowledge cutoff: April 20, 2026
- Cached input pricing (90% discount)
- Adjustable reasoning effort (low to max)
- Function calling and native vision support

## Use Cases
- Frontier reasoning and analysis
- Complex agentic workflows with tool use
- Large-codebase understanding and refactoring
- Long document processing up to 1M tokens
- Repeated context with cached inputs (RAG, codebases)

## API Endpoint

Base URL: https://api.core.today
Endpoint: POST /llm/openai/v1/chat/completions

## Authentication
Header: Authorization: Bearer YOUR_API_KEY

Note: LLM endpoints use OpenAI-compatible format with Authorization Bearer token.

## Input Parameters

### Required
- **messages**: array - Array of message objects with role and content
- **model**: string (default: gpt-6-sol) - Model identifier

### Optional
- **max_completion_tokens**: integer (default: 4096) - Maximum tokens in response (up to 128000). Note: use max_completion_tokens, not max_tokens
- **reasoning_effort**: string (default: medium) - Reasoning effort level: low, medium, high, xhigh, or max
  Options: low, medium, high, xhigh, max
- **temperature**: float (default: 1.0) - Sampling temperature (0-2)
- **stream**: boolean (default: false) - Enable Server-Sent Events streaming

## Examples

### Frontier Agentic Coding
Multi-step code reasoning with GPT-6 Sol
```json
{
  "model": "gpt-6-sol",
  "messages": [
    {
      "role": "system",
      "content": "You are a senior software engineer. Think step by step."
    },
    {
      "role": "user",
      "content": "Design a migration plan from a monolithic Express API to modular services, then generate the first service's code with tests."
    }
  ],
  "reasoning_effort": "high",
  "max_completion_tokens": 8000
}
```

### Cached Repeated Context
Reuse a large system prompt with 90% cached input discount
```json
{
  "model": "gpt-6-sol",
  "messages": [
    {
      "role": "system",
      "content": "<large repeated system prompt or codebase context>"
    },
    {
      "role": "user",
      "content": "Summarize the open TODOs and rank them by risk."
    }
  ],
  "temperature": 0.3,
  "max_completion_tokens": 4000
}
```

## Response Format
```json
{
  "id": "chatcmpl-abc123",
  "object": "chat.completion",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Response text here"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 100,
    "completion_tokens": 50,
    "total_tokens": 150
  }
}
```

## Tips
- Sol ($2/$10 per M) covers most workloads — escalate to GPT-6 Astra ($10/$50) only for the hardest tasks, drop to GPT-6 Luna ($0.10/$0.50) for simple, high-volume calls
- Use cached inputs ($0.20/M, 90% off) for repeated system prompts and RAG context
- Keep prompts under 272K input tokens where possible — above that the whole request bills at 2x input / 1.5x output
- reasoning_effort 'high', 'xhigh' or 'max' for the most complex tasks
- 128K output enables single-shot long-form generation
- Lower temperature (0.2-0.5) for coding and analytical tasks
- Cache writes bill at 1.25x the input rate — implicit caching writes by default, so most of a large first prompt is billed as a cache write, then reads are 90% off
- Fast mode (service_tier: fast or priority) bills token charges at 2x — it is settled on the tier the response reports, so a request downgraded to standard is billed at the standard rate

## Documentation
https://platform.openai.com/docs/models