Model / Commands (reference)
Every model family in TPEE's catalogue has its own command sheet — the slash commands and CLI flags that run outside TPEE, in the model vendor's own tool (Claude Code, Codex, Gemini CLI, Kimi CLI, qwen, DeepCode, Ollama…). Pick a family below to see its sheet exactly as it ships with the app (Build 294, reference as of 2026-08-11). Looking for context windows instead? See the model context & character limits table.
Claude (Anthropic)
Slash commands for Claude Code and the Claude Agent SDK — /help, /plan, /diff, /undo, /review, /compact, /think, /critique, /refactor, /agents and more, including the Opus 5 /patch and /grind additions.
GPT (OpenAI)
Codex CLI commands (codex, codex run, codex apply, codex revert, codex review) plus the in-session slash menu and Responses API hints, with the July 2026 GPT-5.6 price cuts noted.
Gemini (Google)
Gemini CLI commands (gemini, gemini run, gemini --model, gemini auth) and in-session slashes — /think, /tools, /files, /image — plus Vertex AI / AI Studio programmatic hints.
Kimi (Moonshot AI)
Kimi CLI flags and sessions (kimi --yolo, --model, --think, --skills-dir), the full interactive slash command list (/model, /sessions, /skill:…, /flow:…, /yolo, /compact) and the OpenAI-compatible API hints.
DeepSeek
DeepSeek chat and the official DeepCode CLI (deep-thinking + Agent Skills), the model list (V4-Pro, V4-Flash-0731 GA) and the in-chat slashes — /help, /think, /file, /code.
Qwen (Alibaba)
Qwen Code CLI (binary qwen) with qwen3.8-max and the Qwen3-Coder family, in-session slashes (/file, /dir, /diff, /apply) and DashScope API parameters.
Mistral
Mistral Vibe CLI (renamed from the mistral CLI in 2026), Codestral fill-in-the-middle, the model list (Medium 3.5, Large 3, Small 4, Codestral, Devstral) and Le Chat notes.
Grok (xAI)
Grok Build CLI commands (plan/review mode, subagents, headless -p) and the Grok 4.5 / 4.3 model notes, plus the in-session /think and /web slashes.
Llama (Meta)
Open-weights family — run it locally with Ollama (ollama run llama4:scout), llama.cpp (llama-cli) or vLLM; the sheet covers the local runtimes and their commands.
Sonar (Perplexity)
Perplexity's Sonar models with built-in web search — /focus, /image, /file, /share in the app, plus the API's return_citations and search_mode parameters.
Command (Cohere)
Cohere Command family via the v2 API — /v2/chat with tools and documents (native RAG), plus rerank and embed endpoints.
MiniMax
MiniMax M-series (M3, M2.5) via the OpenAI-compatible API and the mmx multimodal CLI — text, image, video, music and speech.
Nova (Amazon)
Amazon Nova on Bedrock — aws bedrock list-foundation-models, the Converse API and the Bedrock Agents / Knowledge Bases console flow. Nova runs from a terminal via the Amazon Q CLI (q).
GLM (Z.AI)
Z.AI GLM family (GLM-5.2 flagship) — the GLM Coding Plan consumed through Claude Code / Codex / Cline with a base-URL swap, plus the OpenAI-compatible z.ai API.
Hermes (Nous Research)
Nous Research Hermes family (Hermes 4 / 4.3 / 3) — the hermes agent CLI, OpenRouter API routes and fully local Ollama / llama.cpp / vLLM serving.
The command sheets are updated weekly in the real app — download TPEE to always carry the latest reference with you, fully offline. TPEE itself never launches these CLIs and never connects to any vendor API or MCP service — the sheets are static reference text only.