Frontier models in one agentic workflow.

Claude Opus 5 for complex agentic coding and reasoning workflows in AdaL.

Claude Sonnet 5 is the most capable Sonnet for coding and agents. Available in AdaL CLI and Desktop.

Claude Sonnet 4.6 for fast, balanced coding and engineering execution in AdaL.

Claude Opus 4.8 for advanced reasoning and complex engineering workflows in AdaL.

Claude Opus 4.7 for advanced reasoning and high-stakes coding workflows in AdaL.

Claude Opus 4.6 for deep reasoning and production code in AdaL. Recommended for complex engineering.

Claude Haiku 4.5 — fast, affordable thinking model for quick coding tasks in AdaL.

GPT-5.6 Terra balances intelligence and cost. The default model in AdaL for everyday coding.

GPT-5.6 Luna for cost-sensitive high-volume workloads in AdaL.

GPT-5.6 Sol — OpenAI frontier model for complex professional work in AdaL.

GPT-5.5 for coding, computer use, and deep research workflows in AdaL.

GPT-5.4 for native computer use, tool search, and large-context workflows in AdaL.

GPT-5.3 Codex — code-specialized reasoning model in AdaL.

GPT-5.2 — general-purpose reasoning model in AdaL.

GPT-5 Mini — the cheapest OpenAI reasoning model in AdaL.

Gemini 3.7 Flash — Google's most capable Flash model for coding and agentic workflows, with 1M context in AdaL.

Gemini 3.1 Pro for multimodal reasoning and frontier-model coding workflows in AdaL.

Gemini 3 Flash — fast, affordable thinking model with 1M context in AdaL.

Gemini 3.6 Flash — token-efficient agentic model with 1M context in AdaL.

Gemini 3.5 Flash — balanced speed and capability with 1M context in AdaL.

MiniMax M2.5 Highspeed — fast, affordable thinking model in AdaL.

MiniMax M2.7 for efficient frontier-model work inside AdaL workflows.

MiniMax M2.7 Highspeed — fast variant of M2.7 in AdaL.

MiniMax M3 — recommended for browser-use, coding, and agentic workflows in AdaL.

Muse Spark 1.1 — Meta's fast agentic coding model with 1M context in AdaL.

Run Meta's Muse Spark 1.2 coding model in AdaL: 1M context, vision, configurable reasoning effort, $1.25/M input and $4.25/M output. Pick it with /model.

Grok 4.6 — xAI's strongest agentic coding model in AdaL.

Grok 4.5 — xAI's strong agentic coding model in AdaL.

Z.ai GLM-5.2 for 1M-token long-horizon coding, flexible effort levels, and open-source weights in AdaL.

Z.ai GLM-5.1 — capable thinking model with flexible effort in AdaL.

Z.ai GLM-5 — affordable thinking model in AdaL.

Z.ai GLM-5-Turbo — fast thinking model in AdaL.

Z.ai GLM-4.7 — affordable thinking model in AdaL.

Z.ai GLM-4.7 FlashX — ultra-fast, ultra-cheap model in AdaL.

Z.ai GLM-4.7 Flash — free thinking model in AdaL.

Z.ai GLM-4.5 Flash — free fast model in AdaL.

DeepSeek V4 Flash — ultra-affordable thinking model in AdaL.

DeepSeek V4 Pro — powerful reasoning model at affordable pricing in AdaL.

Kimi K2.6 — capable thinking model for coding in AdaL.

Kimi K3 — frontier coding model with deep reasoning and 1M context in AdaL.

Kimi K2.7 Code for research-heavy coding, long-context reasoning, and agentic engineering workflows in AdaL.

Claude Fable 5 — next-gen Anthropic flagship for knowledge work and coding in AdaL.

Run Alibaba's Qwen3.8 Max in AdaL: multimodal long-context flagship with a 991K input window, 131K output, and $2/M input · $6/M output. Pick it with /model.

Qwen3.7 Max — high-capability thinking model with ~1M context in AdaL.

Qwen3.7 Plus — affordable thinking model with vision and ~1M context in AdaL.