GPT-6
Generate responses through Frevana's OpenAI Responses API (POST /openai/v1/responses) using gpt-6.
Return the validated API response JSON unchanged, or use --text-only when the user requests only the final text output.
Model Characteristics
- Model ID:
gpt-6 (OpenAI GPT-6 next-generation frontier model).
- Best suited for state-of-the-art multimodal understanding, breakthrough scientific reasoning, complex codebase refactoring, autonomous problem solving, and highest-tier creative synthesis.
- For high-speed/cost-sensitive tasks, consider
gpt-5.6-luna. For standard frontier reasoning, consider gpt-5.6-sol or gpt-5.6-terra.
What This Skill Needs
input or prompt via --input or --input-file (or raw JSON payload via --raw-payload-file)
- optional developer system instructions via
--instructions or --instructions-file
- optional reasoning effort:
--reasoning-effort <low|medium|high>
- optional sampling:
--temperature, --top-p, --max-output-tokens
- optional multi-turn conversation continuation:
--previous-response-id
- optional tools:
--tools or --tools-file, --tool-choice
- optional service tier:
--service-tier <auto|default|flex|priority|ultrafast>
- authentication:
FREVANA_TOKEN (or --token) / FREVANA_API_KEY (or --api-key)
- optional attribution:
FREVANA_AGENT_APP_INSTANCE_ID (or X_FREVANA_AGENT_APP_INSTANCE_ID, or --agent-app-instance-id) sent as x-frevana-agent-app-instance-id
curl, bash, and python3
Allowed Options
--model: gpt-6 (default)
--input, --prompt: user prompt string or JSON message array
--input-file: file path containing input prompt or JSON message array
--instructions: system developer instructions
--instructions-file: file path containing system instructions
--reasoning-effort: low, medium, high
--temperature: 0.0 - 2.0
--top-p: 0.0 - 1.0
--max-output-tokens: positive integer
--session-file: path to session file to automatically save and restore conversation context across calls
--previous-response-id: ID of a previous response for multi-turn conversations
--chat: start an interactive chat session in the terminal
--service-tier: auto, default, flex, priority, ultrafast
--tools: JSON array string defining available tools/functions
--tools-file: path to JSON file defining tools
--tool-choice: tool choice behavior (auto, required, none, or JSON)
--raw-payload-file: path to raw JSON request body
--api-key: Frevana API key (sent as X-API-Key)
--token: Frevana Bearer token (sent as Authorization: Bearer)
--agent-app-instance-id: Agent App instance/team ID (sent as x-frevana-agent-app-instance-id)
--text-only, -t: output only the response text
--output: save full response JSON to path
Commands
# Frontier reasoning with GPT-6
bash <skill-path>/scripts/create_response.sh \
--input "Formulate a formal proof for safety invariants in an asynchronous Paxos consensus implementation" \
--instructions "Provide rigorous mathematical definitions and complete inductive verification steps." \
--reasoning-effort high \
--output ./out/gpt6-response.json
# Multi-turn conversation with automatic session persistence
bash <skill-path>/scripts/create_response.sh \
--session-file ./out/gpt6_session.json \
--input "Analyze the security vulnerabilities in this smart contract architecture" \
--text-only
bash <skill-path>/scripts/create_response.sh \
--session-file ./out/gpt6_session.json \
--input "Provide a formal patch for the reentrancy issue identified above" \
--text-only
# Interactive terminal chat
bash <skill-path>/scripts/create_response.sh --chat
1---2name: gpt-63description: Use when the user wants next-generation frontier intelligence, state-of-the-art reasoning, groundbreaking multimodal synthesis, or advanced cognitive workflows using OpenAI GPT-6 (gpt-6) through the Frevana Responses API.4metadata:5 source: https://skillmd.com/skills/finpeakinc/gpt-66---78# GPT-6910Generate responses through Frevana's OpenAI Responses API (`POST /openai/v1/responses`) using `gpt-6`.1112Return the validated API response JSON unchanged, or use `--text-only` when the user requests only the final text output.1314## Model Characteristics1516- Model ID: `gpt-6` (OpenAI GPT-6 next-generation frontier model).17- Best suited for state-of-the-art multimodal understanding, breakthrough scientific reasoning, complex codebase refactoring, autonomous problem solving, and highest-tier creative synthesis.18- For high-speed/cost-sensitive tasks, consider `gpt-5.6-luna`. For standard frontier reasoning, consider `gpt-5.6-sol` or `gpt-5.6-terra`.1920## What This Skill Needs2122- `input` or `prompt` via `--input` or `--input-file` (or raw JSON payload via `--raw-payload-file`)23- optional developer system instructions via `--instructions` or `--instructions-file`24- optional reasoning effort: `--reasoning-effort <low|medium|high>`25- optional sampling: `--temperature`, `--top-p`, `--max-output-tokens`26- optional multi-turn conversation continuation: `--previous-response-id`27- optional tools: `--tools` or `--tools-file`, `--tool-choice`28- optional service tier: `--service-tier <auto|default|flex|priority|ultrafast>`29- authentication: `FREVANA_TOKEN` (or `--token`) / `FREVANA_API_KEY` (or `--api-key`)30- optional attribution: `FREVANA_AGENT_APP_INSTANCE_ID` (or `X_FREVANA_AGENT_APP_INSTANCE_ID`, or `--agent-app-instance-id`) sent as `x-frevana-agent-app-instance-id`31- `curl`, `bash`, and `python3`3233## Allowed Options3435- `--model`: `gpt-6` (default)36- `--input`, `--prompt`: user prompt string or JSON message array37- `--input-file`: file path containing input prompt or JSON message array38- `--instructions`: system developer instructions39- `--instructions-file`: file path containing system instructions40- `--reasoning-effort`: `low`, `medium`, `high`41- `--temperature`: `0.0 - 2.0`42- `--top-p`: `0.0 - 1.0`43- `--max-output-tokens`: positive integer44- `--session-file`: path to session file to automatically save and restore conversation context across calls45- `--previous-response-id`: ID of a previous response for multi-turn conversations46- `--chat`: start an interactive chat session in the terminal47- `--service-tier`: `auto`, `default`, `flex`, `priority`, `ultrafast`48- `--tools`: JSON array string defining available tools/functions49- `--tools-file`: path to JSON file defining tools50- `--tool-choice`: tool choice behavior (`auto`, `required`, `none`, or JSON)51- `--raw-payload-file`: path to raw JSON request body52- `--api-key`: Frevana API key (sent as `X-API-Key`)53- `--token`: Frevana Bearer token (sent as `Authorization: Bearer`)54- `--agent-app-instance-id`: Agent App instance/team ID (sent as `x-frevana-agent-app-instance-id`)55- `--text-only`, `-t`: output only the response text56- `--output`: save full response JSON to path5758## Commands5960```bash61# Frontier reasoning with GPT-662bash <skill-path>/scripts/create_response.sh \63 --input "Formulate a formal proof for safety invariants in an asynchronous Paxos consensus implementation" \64 --instructions "Provide rigorous mathematical definitions and complete inductive verification steps." \65 --reasoning-effort high \66 --output ./out/gpt6-response.json6768# Multi-turn conversation with automatic session persistence69bash <skill-path>/scripts/create_response.sh \70 --session-file ./out/gpt6_session.json \71 --input "Analyze the security vulnerabilities in this smart contract architecture" \72 --text-only7374bash <skill-path>/scripts/create_response.sh \75 --session-file ./out/gpt6_session.json \76 --input "Provide a formal patch for the reentrancy issue identified above" \77 --text-only7879# Interactive terminal chat80bash <skill-path>/scripts/create_response.sh --chat81```