Cmd. /model

Purpose

Switch among GPT-6 Astra, GPT-6.1 Sol, and GPT-6 Luna, the earlier GPT-6 Sol and GPT-5.6 models, and supported reasoning effort levels.

/model

Availability

The installed command menu and account or organization settings determine availability.

Examples

/model

When to use it

1. Select GPT-6 Astra with high~max reasoning before the hardest and longest work.
2. Switch to GPT-6.1 Sol for everyday and complex coding or agentic work.
3. Use GPT-6 Luna for clear, repeatable tasks such as summarization or extraction.
4. Run /status afterward to confirm the model, reasoning, and approval policy.

Choosing related commands

Effects and persistence

Uses the selected model for subsequent requests. Access, reasoning settings, and pricing depend on the chosen model.

/model chooses the active model and, in supported flows, reasoning effort for the current Codex session.

GPT-6.1 Sol at a Glance

The rollout to Codex and ChatGPT Work began on September 29, 2026. The official Models page describes it as near-Astra performance at a lower cost and recommends it for complex coding and agentic workflows.

Standard and Fast modes are available, and Ultrafast is available on Pro $500 and eligible Enterprise and Edu plans. In Codex, reasoning ranges from Light to Ultra; Max and Ultra depend on your settings.

model = "gpt-6.1-sol"

GPT-6 Astra at a Glance

model = "gpt-6-astra"
model_reasoning_effort = "high"

The 0.159.1 default change (GPT-6 Astra to GPT-6.1 Sol) applies only to sessions without an explicit model in config.toml, so teams that pin a standard are not affected. Astra also appears in the Amazon Bedrock Mantle and Runtime pickers from 0.153.3, and GPT-6.1 Sol is the default in both catalogs from 0.159.1.

GPT-6 Sol and Luna at a Glance

Both joined the model picker in 0.156.1; 0.157.0 added them to Amazon Bedrock catalogs and added migration prompts for older models. The rate-limit switch prompt recommends GPT-6 Luna.

GPT-5.5 retired from ChatGPT-authenticated Codex on 2026-10-14 at 19:00 UTC. The OpenAI API and API-key authenticated sessions are unaffected. Check /model for a plan/client-available replacement such as GPT-6.1 Sol (gpt-6.1-sol) and update workspace defaults, saved settings, managed configuration, custom agents, scheduled tasks, and scripts referencing gpt-5.5. Official retirement notice.

model = "gpt-6.1-sol"

Current Selection Notes

• Top model and hardest work: GPT-6 Astra (gpt-6-astra)
• Everyday and complex coding or agentic work: GPT-6.1 Sol (gpt-6.1-sol, bundled default from CLI 0.159.1)
• Earlier Sol: GPT-6 Sol (gpt-6-sol, picker from CLI 0.156.1) remains selectable
• Clear repeated or cost-sensitive work: GPT-6 Luna (gpt-6-luna, picker from CLI 0.156.1)
• During the rollout: GPT-5.6 Sol, Terra, and Luna remain selectable
• GPT-5.5 retired from ChatGPT-authenticated Codex on 2026-10-14. API-key sessions are unaffected; choose an available replacement such as GPT-6.1 Sol.
• Retired: GPT-5.3-Codex-Spark on 2026-09-14; GPT-5.4 and GPT-5.4 mini left ChatGPT-sign-in Codex on 2026-08-31
• Remove from new defaults: gpt-5.2 and gpt-5.3-codex for ChatGPT sign-in

GPT-5.5 retired from ChatGPT-authenticated Codex on 2026-10-14. API-key sessions are unaffected; choose an available replacement such as GPT-6.1 Sol.

Cautions

• Use the account's /model picker for exact IDs and reasoning levels. Selecting a new model does not grant workspace permissions or access.
• If GPT-6.1 Sol is not listed in the picker, choose another model. On Enterprise and Edu, an administrator must enable it first.
• Bound concurrency, write scope, and credits before using Ultra delegation.
• Codex Cloud tasks currently use a fixed default model.
• Pair high-risk model choices with /permissions and human review.

Official sources

• Developer Commands
• Models
• Changelog
• Rust V0.157.0
• GPT-6.1 Sol
• Rust V0.159.1

Complete command list · Verification scope · History for this command