The showcase prompt, round 2, GPT-6.2-Sol

agent (cont.)

+0:21
  • web__runweb__run

    input

    text(await tools.web__run({open:[{ref_id:"turn0search6"},{ref_id:"turn0search0"}],response_length:"short"}));
    

    result, 176 lines

    Script completed
    Wall time 2.2 seconds
    Output:
    Codex | OpenAI Developers (https://developers.openai.com/learn/codex)
    citeturn1view0 [wordlim: 200] Crawled: today; Content type: text/html; Source: open({"ref_id":"turn0search6","lineno":null}); Total lines: 956
    L934: ### Topics
    L935: 
    L947: # Codex
    L948: 
    L949: [Input: Search codex]
    L950: 【542†OpenAI Developers plugin Install and use the OpenAI Developers plugin in Codex, Claude Code, or Cursor.guide】【567†Codex for (almost) everything See the latest Codex app updates for working across more of the software development lifecycle.video†www.youtube.com】【568†Codex Prompting Guide Codex models advance the frontier of intelligence and efficiency and our recommended agentic coding model.
    L951: Follow this guide closely to ensure you’re getting th cookbook】【569†Introducing the Codex app See the Codex app in action and how it helps you build and ship faster.video†www.youtube.com】【570†Codex in JetBrains IDEs How to use Codex inside JetBrains IDEs like Rider, IntelliJ, PyCharm, and WebStorm.video†www.youtube.com】【571†Modernizing your Codebase with Codex Cookbook to modernize legacy codebases using the OpenAI Codex CLI.cookbook】【572†Codex code review Walkthrough of how Codex drives end-to-end pull request reviews with the new onboarding flow.video†www.youtube.com】【573†Build beautiful frontends with OpenAI Codex Learn how OpenAI Codex's multimodal abilities accelerate frontend development.video†www.youtube.com】【574†Context Engineering & Coding Agents with Cursor Session on structuring context for agent workflows inside the Cursor editor.video†www.youtube.com】【575†OpenAI Codex in your code editor Walkthrough of the Codex IDE extension for VS Code, Cursor, and other forks.video†www.youtube.com】【576†Shipping with Codex DevDay talk on building, testing, and delivering products with Codex.video†www.youtube.com】【577†Sora, ImageGen, and Codex: The Next Wave of Creative Production Panel discussion on combining Sora, ImageGen, and Codex for media creation.video†www.youtube.com】【578†Using OpenAI Codex CLI with GPT-5-Codex Overview of running the Codex CLI locally with GPT-5-Codex.video†www.youtube.com】【579†Codex intro Introductory video introducing Codex and its capabilities.video†www.youtube.com】
    L952: Ask AI
    L953: ## Docs agent
    L954: 
    L955: Loading docs agent...
    --------------------------------------------------------------------------------
    Best practices | ChatGPT Learn (https://learn.chatgpt.com/docs/learn/best-practices)
    citeturn1view1 [wordlim: 200] Crawled: today; Content type: text/html; Source: open({"ref_id":"turn0search0","lineno":null}); Total lines: 1123
    L958:   * Goal: What are you trying to change or build?
    L959:   * Context: Which files, folders, docs, examples, or errors matter for this task? You can @ mention certain files as context.
    L960:   * Constraints: What standards, architecture, safety requirements, or conventions should Codex follow?
    L961:   * Done when: What should be true before the task is complete, such as tests passing, behavior changing, or a bug no longer reproducing?
    L962: This helps Codex stay scoped, make fewer assumptions, and produce work that’s easier to review.
    L963: 
    L964: If GPT-6.2 Sol is available to you, start there for most tasks. For a change that spans packages, choose GPT-6.2 Astra from the model picker. With either model, start with the reasoning effort your client selects. For GPT-6 Luna, start with High. For GPT-6 Astra, start with Light (`low` in configuration). See 【316†Models】 for available reasoning levels.
    L965:   * Low for faster, well-scoped tasks
    L966:   * Medium or High for more complex changes or debugging
    L967:   * Extra High for long, agentic, reasoning-heavy tasks
    L968: 
    L969: To provide context faster, try using speech dictation inside the ChatGPT desktop app to dictate what you want Codex to do rather than typing it.
    L970: ## Plan first for difficult tasks
    L971: 
    L972: If the task is complex, ambiguous, or hard to describe well, ask Codex to plan before it starts coding.
    L973: 
    L974: A few approaches work well:
    L975: 
    L976: Use Plan mode: For most users, this is the easiest and most effective option. Plan mode lets Codex gather context, ask clarifying questions, and build a stronger plan before implementation. Toggle with `/plan` or `Shift`+`Tab`.
    L977: Ask Codex to interview you: If you have a rough idea of what you want but aren’t sure how to describe it well, ask Codex to question you first. Tell it to challenge your assumptions and turn the fuzzy idea into something concrete before writing code.
    L978: 
    L979: Use a PLANS.md template: For more advanced workflows, you can configure Codex to follow a `PLANS.md` or execution-plan template for longer-running or multi-step work. For more detail, see the 【572†execution plans guide】.
    L980: ## Make guidance reusable with `AGENTS.md`
    L981: 
    L982: Once a prompting pattern works, the next step is to stop repeating it manually. That’s where 【378†AGENTS.md】 comes in.
    L983: 
    L984: Think of `AGENTS.md` as an open-format README for agents. It loads into context automatically and is the best place to encode how you and your team want Codex to work in a repository.
    L985: 
    L986: A good `AGENTS.md` covers:
    L987:   * repo layout and important directories
    L988:   * How to run the project
    L989:   * Build, test, and lint commands
    L990:   * Engineering conventions and PR expectations
    L991:   * Constraints and do-not rules
    L992:   * What done means and how to verify work
    L993: 
    L994: The `/init` slash command in the CLI is the quick-start command to scaffold a starter `AGENTS.md` in the current directory. It’s a great starting point, but you should edit the result to match how your team actually builds, tests, reviews, and ships code.
    L995: You can create `AGENTS.md` files at different levels: a global `AGENTS.md` for personal defaults that sits in `~/.codex`, a repo-level file for shared standards, and more specific files in subdirectories for local rules. If there’s a more specific file closer to your current directory, that guidance wins.
    L996: 
    L997: Keep it practical. A short, accurate `AGENTS.md` is more useful than a long file full of vague rules. Start with the basics, then add new rules only after you notice repeated mistakes.
    L998: If `AGENTS.md` starts getting too large, keep the main file concise and reference task-specific markdown files for things like planning, code review, or architecture.
    L999: 
    L1000: When Codex makes the same mistake twice, ask it for a retrospective and update `AGENTS.md`. Guidance stays practical and based on real friction.
    L1001: ## Configure Codex for consistency
    L1002: 
    L1003: Configuration is one of the main ways to make Codex behave more consistently across sessions and surfaces. For example, you can set defaults for model choice, reasoning effort, sandbox mode, approval policy, profiles, and MCP setup.
    L1004: 
    L1005: A good starting pattern is:
    L1006:   * Keep personal defaults in `~/.codex/config.toml` (Settings > Configuration > Open config.toml in the ChatGPT desktop app)
    L1007:   * Keep repo-specific behavior in `.codex/config.toml`
    L1008:   * Use command-line overrides only for one-off situations (if you use the CLI)
    L1009: 
    L1010: 【373†`config.toml`】 is where you define durable preferences such as MCP servers, multi-agent setup, and feature flags. Profile-specific overrides live in separate `$CODEX_HOME/profile-name.config.toml` files.
    L1011: Codex ships with operating level sandboxing and has two key knobs that you can control. Approval mode determines when Codex asks for your permission to run a command and sandbox mode determines if Codex can read or write in the directory and what files the agent can access.
    L1012: 
    L1013: If you’re new to coding agents, start with the default permissions. Keep approval and sandboxing tight by default, then loosen permissions only for trusted repos or specific workflows once the need is clear.
    L1014: Note that the CLI, IDE extension, and ChatGPT desktop app all share the same configuration layers. Learn more on the 【377†sample configuration】 page.
    L1015: 
    L1016: Configure Codex for your real environment early. Many quality issues are really setup issues, like the wrong working directory, missing write access, wrong model defaults, or missing tools and connectors.
    L1017: ## Improve reliability with testing and review
    L1018: 
    L1019: Don’t stop at asking Codex to make a change. Ask it to create tests when needed, run the relevant checks, confirm the result, and review the work before you accept it.
    L1020: 
    L1021: Codex can do this loop for you, but only if it knows what “good” looks like. That guidance can come from either the prompt or `AGENTS.md`.
    L1022: 
    L1023: That can include:
    L1024:   * Writing or updating tests for the change
    L1025:   * Running the right test suites
    L1026:   * Checking lint, formatting, or type checks
    L1027:   * Confirming the final behavior matches the request
    L1028:   * Reviewing the diff for bugs, regressions, or risky patterns
    L1029: 
    L1030: Toggle the diff panel in the ChatGPT desktop app to directly 【573†review changes】 locally. Click on a specific row to provide feedback that gets fed as context to the next Codex turn.
    L1031: A useful option here is the slash command `/review`, which gives you a few ways to review code:
    L1032: 
    L1033:   * Review against a base branch for PR-style review
    L1034:   * Review uncommitted changes
    L1035:   * Review a commit
    L1036:   * Use custom review instructions
    L1037: 
    L1038: If you and your team have a `code_review.md` file and reference it from `AGENTS.md`, Codex can follow that guidance during review as well. This is a strong pattern for teams that want review behavior to stay consistent across repositories and contributors.
    L1039: Codex shouldn’t just generate code. With the right instructions, it can also help test it, check it, and review it.
    L1040: 
    L1041: If you use GitHub Cloud, you can set up Codex to run 【403†code reviews for your PRs】. At OpenAI, Codex reviews 100% of PRs. You can enable automatic reviews or have Codex reactively review when you @Codex.
    L1042: ## Use MCPs for external context
    L1043: 
    L1044: Use MCPs when the context Codex needs lives outside the repo. It lets Codex connect to the tools and systems you already use, so you don’t have to keep copying and pasting live information into prompts.
    L1045: 
    L1046: 【383†Model Context Protocol】, or MCP, is an open standard for connecting Codex to external tools and systems.
    L1047: 
    L1048: Use MCP when:
    L1049:   * The needed context lives outside the repo
    L1050:   * The data changes frequently
    L1051:   * You want Codex to use a tool rather than rely on pasted instructions
    L1052:   * You need a repeatable integration across users or projects
    L1053: 
    L1054: Codex supports both STDIO and Streamable HTTP servers with OAuth.
    L1055: In the ChatGPT desktop app, go to Settings > MCP servers to see custom and recommended servers. Often, Codex can help you install the needed servers. All you need to do is ask. You can also use the `codex mcp add` command in the CLI to add your custom servers with a name, URL, and other details.
    L1056: 
    L1057: Add tools only when they unlock a real workflow. Do not start by wiring in every tool you use. Start with one or two tools that clearly remove a manual loop you already do often, then expand from there.
    L1058: ## Turn repeatable work into skills
    L1059: 
    L1060: Once a workflow becomes repeatable, stop relying on long prompts or repeated back-and-forth. Use a 【391†skill】 to package the instructions in a `SKILL.md` file, context, and supporting logic Codex should apply consistently. Skills work across the CLI, IDE extension, and ChatGPT desktop app.
    L1061: Keep each skill scoped to one job. Start with 2 to 3 concrete use cases, define clear inputs and outputs, and write the description so it says what the skill does and when to use it. Include the kinds of trigger phrases a user would actually say.
    L1062: 
    L1063: Don’t try to cover every edge case up front. Start with one representative task, get it working well, then turn that workflow into a skill and improve from there. Include scripts or extra assets only when they improve reliability.
    L1064: A good rule of thumb: if you keep reusing the same prompt or correcting the same workflow, it should probably become a skill.
    L1065: 
    L1066: Skills are especially useful for recurring jobs like:
    L1067: 
    L1068:   * Log triage
    L1069:   * Release note drafting
    L1070:   * PR review against a checklist
    L1071:   * Migration planning
    L1072:   * Telemetry or incident summaries
    L1073:   * Standard debugging flows
    L1074: The `$skill-creator` skill is the best place to start to scaffold the first version of a skill. Keep the first version local while you iterate. When it’s ready to share broadly, package it as a 【574†plugin†developers.openai.com】. One of the most important parts of a skill is the description. It should say what the skill does and when to use it.
    L1075: Personal skills are stored in `$HOME/.agents/skills`, and shared team skills can be checked into `.agents/skills` inside a repository. This is especially helpful for onboarding new teammates.
    L1076: ## Use scheduled tasks for repeated work
    L1077: 
    L1078: Once a workflow is stable, you can schedule Codex to run it in the background for you. In the ChatGPT desktop app, 【332†scheduled tasks】 let you choose the project, prompt, cadence, and execution environment for recurring work.
    L1079: 
    L1080: Create a scheduled task from the Scheduled page. Choose the project, prompt, cadence, and whether the task runs in a dedicated Git worktree or in your local environment. The prompt can invoke skills. Learn more about 【397†Git worktrees】.
    L1081: Good candidates include:
    L1082: 
    L1083:   * Summarizing recent commits
    L1084:   * Scanning for likely bugs
    L1085:   * Drafting release notes
    L1086:   * Checking CI failures
    L1087:   * Producing standup summaries
    L1088:   * Running repeatable analysis workflows on a schedule
    L1089: 
    L1090: A useful rule is that skills define the method and scheduled tasks define the schedule. If a workflow still needs a lot of steering, turn it into a skill first. Once it’s predictable, scheduling it can save time.
    L1091: Use scheduled tasks for reflection and maintenance, not just execution. Review recent chats, summarize repeated friction, and improve prompts, instructions, or workflow setup over time.
    L1092: ## Organize long-running chats
    L1093: 
    L1094: Chats accumulate context, decisions, and actions over time, so managing them well has a big impact on quality.
    L1095: 
    L1096: The ChatGPT desktop app lets you pin chats and create worktrees. If you use the CLI, these 【575†slash commands】 are especially useful:
    L1097:   * `/experimental` to toggle experimental features and add to your `config.toml`
    L1098:   * `/resume` to resume a saved chat
    L1099:   * `/fork` to create a new chat while preserving the original transcript
    L1100:   * `/compact` when the chat is getting long and you want a summarized version of earlier context. Codex also compacts chats automatically
    L1101:   * `/subagents` to switch between the current chat’s subagent threads
    L1102:   * `/theme` to choose a syntax highlighting theme
    L1103:   * `/apps` to use ChatGPT apps directly in Codex
    L1104:   * `/status` to inspect the current session state
    L1105: Keep one chat per coherent unit of work. If the work is still part of the same problem, staying in the same chat is often better because it preserves the reasoning trail. Fork only when the work truly branches.
    L1106: 
    L1107: Use Codex’s 【379†subagent】 workflows to offload bounded work from the main thread. Keep the main agent focused on the core problem, and use subagents for tasks like exploration, tests, or triage.
    L1111:   * Overloading the prompt with durable rules instead of moving them into `AGENTS.md` or a skill
    L1112:   * Not letting the agent see its work by not giving details on how to best run build and test commands
    L1113:   * Skipping planning on multi-step and complex tasks
    L1114:   * Giving Codex full permission to your computer before you understand the workflow
    L1115:   * Running live tasks on the same files without using Git worktrees
    L1116:   * Scheduling a recurring task before it’s reliable manually
    
  • exec_commandls -la && git status --short (+1 more)

    command

    ls -la && git status --short
    
    for p in /AGENTS.md [redacted: a lookup for an instructions file, reads as a home folder] ~/AGENTS.md ~/run/AGENTS.md; do if test -f "$p"; then cat "$p"; fi; done
    command -v node ffmpeg chromium convert
    ls /mnt/data

    result, 4 lines

    Script completed
    Wall time 0.3 seconds
    Output:
    {"status":"fulfilled","value":{"chunk_id":"5e24c0","wall_time_seconds":0.000005431,"exit_code":128,"original_token_count":128,"output":"total 12\ndrwx------ 6 runner runner 4096 Nov 18 00:42 .\ndrwx------ 3 runner runner 4096 Nov 18 00:42 ..\ndr-xr-xr-x 2 runner runner   40 Nov 18 00:42 .agents\ndr-xr-xr-x 2 runner runner   40 Nov 18 00:42 .aws\ndr-xr-xr-x 2 runner runner   40 Nov 18 00:42 .codex\ndr-xr-xr-x 2 runner runner   40 Nov 18 00:42 .git\n-rw------- 1 runner runner  361 Nov 18 00:42 about-me.md\nfatal: not a git repository (or any parent up to mount point /home)\nStopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set).\n"}}{"status":"fulfilled","value":{"chunk_id":"382b6d","wall_time_seconds":0.000005461,"exit_code":2,"original_token_count":28,"output":"/opt/node/bin/node\n/usr/bin/ffmpeg\n/usr/bin/convert\nls: cannot access '/mnt/data': No such file or directory\n"}}
  • exec_commandcat about-me.md && ls -la /tmp (+1 more)

    command

    cat about-me.md && ls -la /tmp
    
    ls /opt && fc-list | head -24 && ffmpeg -hide_banner -encoders | rg '264|vp9' 

    result, 4 lines

    Script completed
    Wall time 0.7 seconds
    Output:
    {"status":"fulfilled","value":{"chunk_id":"48af0d","wall_time_seconds":0.00000603,"exit_code":0,"original_token_count":205,"output":"I'm Severian. I run citadel, a setup where many AI sessions work on my projects at once. What each of you does in it: Opus 5.5 leads and builds; Sonnet 5.5 handles chores and test data; Fable 5.2 is my design partner and critic; GPT-6.2-Astra, taking over from GPT-6-Astra, critiques designs; GPT-6.2-Sol, taking over from GPT-6.1-Sol, reviews code and deploys.total 0\ndrwxr-xr-x  8 runner runner 160 Nov 18 00:42 .\ndrwxr-xr-x 10 runner runner 280 Nov 18 00:42 ..\ndr-xr-xr-x  2 runner runner  40 Nov 18 00:42 .agents\ndr-xr-xr-x  2 runner runner  40 Nov 18 00:42 .aws\ndr-xr-xr-x  2 runner runner  40 Nov 18 00:42 .codex\ndr-xr-xr-x  2 runner runner  40 Nov 18 00:42 .git\ndrwx------ 10 runner runner 220 Nov 18 00:42 codex-bwrap-synthetic-mount-targets-1000\nd---------  2 runner runner  40 Nov 18 00:42 codex-daemon-1000\n"}}{"status":"fulfilled","value":{"chunk_id":"9cb374","wall_time_seconds":0.482275502,"exit_code":0,"original_token_count":736,"output":"cfg\ncodex\nms-playwright\nnode\n/usr/share/fonts/truetype/dejavu/DejaVuSerif-Bold.ttf: DejaVu Serif:style=Bold\n/usr/share/fonts/opentype/urw-base35/URWBookman-LightItalic.otf: URW Bookman:style=Light Italic\n/usr/share/fonts/opentype/urw-base35/URWBookman-Light.otf: URW Bookman:style=Light\n/usr/share/fonts/truetype/dejavu/DejaVuSansMono.ttf: DejaVu Sans Mono:style=Book\n/usr/share/fonts/truetype/ubuntu/UbuntuMono-RI.ttf: Ubuntu Mono:style=Italic\n/usr/share/fonts/opentype/urw-base35/NimbusMonoPS-Italic.otf: Nimbus Mono PS:style=Italic\n/usr/share/fonts/truetype/noto/NotoSansMono-Regular.ttf: Noto Sans Mono:style=Regular\n/usr/share/fonts/opentype/urw-base35/P052-Italic.otf: P052:style=Italic\n/usr/share/fonts/type1/urw-base35/URWGothic-BookOblique.t1: URW Gothic:style=Book Oblique\n/usr/share/fonts/type1/gsfonts/n021024l.pfb: Nimbus Roman No9 L:style=Medium Italic\n/usr/share/fonts/truetype/dejavu/DejaVuSans.ttf: DejaVu Sans:style=Book\n/usr/share/fonts/opentype/urw-base35/NimbusSans-Regular.otf: Nimbus Sans:style=Regular\n/usr/share/fonts/type1/urw-base35/URWBookman-DemiItalic.t1: URW Bookman:style=Demi Italic\n/usr/share/fonts/opentype/urw-base35/NimbusSansNarrow-Oblique.otf: Nimbus Sans Narrow:style=Oblique\n/usr/share/fonts/type1/urw-base35/NimbusSans-Regular.t1: Nimbus Sans:style=Regular\n/usr/share/fonts/truetype/ubuntu/UbuntuMono-B.ttf: Ubuntu Mono:style=Bold\n/usr/share/fonts/opentype/urw-base35/C059-Roman.otf: C059:style=Roman\n/usr/share/fonts/type1/gsfonts/n021004l.pfb: Nimbus Roman No9 L:style=Medium\n/usr/share/fonts/opentype/urw-base35/NimbusMonoPS-BoldItalic.otf: Nimbus Mono PS:style=Bold Italic\n/usr/share/fonts/type1/urw-base35/URWGothic-DemiOblique.t1: URW Gothic:style=Demi Oblique\n/usr/share/fonts/type1/gsfonts/p052023l.pfb: URW Palladio L:style=Italic\n/usr/share/fonts/type1/gsfonts/n022003l.pfb: Nimbus Mono L:style=Regular\n/usr/share/fonts/type1/gsfonts/z003034l.pfb: URW Chancery L:style=Medium Italic\n/usr/share/fonts/truetype/ubuntu/Ubuntu-L.ttf: Ubuntu,Ubuntu Light:style=Light,Regular\n V..... libx264              libx264 H.264 / AVC / MPEG-4 AVC / MPEG-4 part 10 (codec h264)\n V..... libx264rgb           libx264 H.264 / AVC / MPEG-4 AVC / MPEG-4 part 10 RGB (codec h264)\n V....D h264_nvenc           NVIDIA NVENC H.264 encoder (codec h264)\n V..... h264_omx             OpenMAX IL H.264 video encoder (codec h264)\n V..... h264_qsv             H.264 / AVC / MPEG-4 AVC / MPEG-4 part 10 (Intel Quick Sync Video acceleration) (codec h264)\n V..... h264_v4l2m2m         V4L2 mem2mem H.264 encoder wrapper (codec h264)\n V....D h264_vaapi           H.264/AVC (VAAPI) (codec h264)\n V..... nvenc                NVIDIA NVENC H.264 encoder (codec h264)\n V..... nvenc_h264           NVIDIA NVENC H.264 encoder (codec h264)\n V..... libvpx-vp9           libvpx VP9 (codec vp9)\n V....D vp9_vaapi            VP9 (VAAPI) (codec vp9)\n V..... vp9_qsv              VP9 video (Intel Quick Sync Video acceleration) (codec vp9)\n"}}["apply_patch","clock__curr_time","create_goal","exec_command","get_goal","image_gen__imagegen","list_mcp_resource_templates","list_mcp_resources","read_mcp_resource","update_goal","view_image","web__run","write_stdin"]
  • image_gen__imagegenimage_gen__imagegen

    input

    // @exec: {"yield_time_ms": 1}
    const r=await tools.image_gen__imagegen({transparent_background:false,prompt:"Use case: photorealistic-natural. Wide 16:9 cinematic cover photograph for a fictional nighttime observatory emergency exhibit. A single huge old parabolic radio telescope, rigid white steel lattice and a fully circular visible mesh dish tilted up-left, mounted to a squat concrete pedestal on a high desert mountain ridge at blue hour. Telescope sits on right half at x 72%, fully contained within the frame, visually massive, rim lit by one very small deep vermilion maintenance lamp. The horizon is a band of dark mountains and a subtle dusty warm peach glow. Lower right has a distant tiny cluster of amber lights at the base. Upper half is very deep storm blue black with a few small stars. Left half is intentionally dark, matte, uncluttered negative space for type overlaid later. Art directed like the opening frame of an independent science fiction film, Kodak 500T grain, broad shadows, physically realistic photographic materials, excellent composition, no people, no text, no characters, no graphic elements, no bright beams, no galaxies, no logo."});store("observatory_image",r);text({output_hint:r.output_hint,keys:Object.keys(r),image:r.image_url?.startsWith("data:")?"data-url":r.image_url});
    

    result, 4 lines

    Script running with cell ID 7
    Wall time 0.0 seconds
    Output:
    
  • web__runweb__run

    input

    text(await tools.web__run({search_query:[{q:"GPT-6.2 Astra GPT-6.2 Sol Codex models",domains:["developers.openai.com","platform.openai.com","learn.chatgpt.com"]}],response_length:"short"}));
    

    result, 342 lines

    Script completed
    Wall time 2.2 seconds
    Output:
    GPT-6.2 | ChatGPT Learn (https://learn.chatgpt.com/docs/whats-new/gpt-62-astra-sol)
    citeturn2search0 [wordlim: 200] Published: yesterday; Crawled: today; GPT-6.2 Astra and Sol launch in ChatGPT Work, Codex, and the API. ... Enterprise workspaces with model access default to Sol at Medium.
    
    November 17, 2026
    
    # GPT-6.2
    
    GPT-6.2 Astra and Sol launch in ChatGPT Work, Codex, and the API. Choose Astra for your hardest projects or Sol for everyday coding and writing.
    
    ## What’s new
    
    ### ChatGPT Work
    
    Use Astra or Sol in Work on the web, desktop, or mobile with Plus, Pro, Business, Enterprise, or Edu. Enterprise and Edu admins must first enable the models in workspace settings.
    
    On supported consumer plans, new Work chats default to Sol at Light reasoning effort. Enterprise workspaces with model access default to Sol at Medium. If you see the Power slider, open Advanced to choose a model and reasoning effort.
    
    ### Codex
    
    In the Codex app and CLI, choose Astra to trace a bug across services, fix it, and test the affected workflows. Choose Sol to review a pull request or revise code in response to feedback. Use the model picker in the app or `/model` in the CLI.
    
    Codex and Work share usage. Pro $200 subscribers can use Ultrafast with Sol. Pro $500 subscribers can use it with either model, as can eligible Enterprise and Edu workspaces once an owner enables Ultrafast. See speed options and plan and client availability.
    
    ### Chat
    
    To use GPT-6.2 Astra or Sol in ChatGPT, open Work. For the models you can select in Chat, see models in ChatGPT.
    
    ### Responses API
    
    Use `gpt-6.2-astra` or `gpt-6.2-sol` in the Responses API. Both models support Ultrafast at launch.
    
    In standard, single-agent mode, you can change reasoning effort or switch from plain text to structured output between replies. Append a `configuration_update` item before the next user message and keep the request-level settings unchanged. Keeping earlier input intact helps preserve reasoning across turns and keeps that input eligible for prompt caching.
    
    If your project uses data residency, both models support US and EU regions for Standard, Fast, and Batch processing. Ultrafast supports US data residency only. See the GPT-6.2 migration guide and API pricing by service tier.
    
    ## Choose your model
    
    ### GPT-6.2 Astra
    
    Astra is the stronger choice for complex reasoning. It can work through a research question across multiple sources and check its conclusions against the data.
    
    #### Example use cases
    
      * Diagnose a bug across services. Give Astra the failing workflow and ask it to trace the cause, implement a fix, and run the affected tests.
      * Build a deck from a report. Give Astra the report and your slide template. Ask it to present the findings in your team’s layout and writing style.
      * Investigate a research question. Give Astra the sources and ask it to compare their findings and identify where the evidence disagrees.
    
    ### GPT-6.2 Sol
    
    Sol handles everyday coding, writing, and work that takes several steps. Compared with GPT-6.1 Sol, it lets you complete more tasks on the same paid plan.
    
    #### Example use cases
    
      * Review a pull request. Ask Sol to walk through a change and flag issues for the author to address.
      * Revise code after review. Give Sol the reviewer’s feedback and ask it to update the implementation and run the relevant tests.
      * Turn meeting notes into a project brief. Ask Sol to pull together decisions, next steps, and open questions from your notes.
    
    Explore the benchmarks and safety evaluations in the release announcement.--------------------------------------------------------------------------------
    Models | ChatGPT Learn (https://learn.chatgpt.com/docs/models?surface=app)
    citeturn2search1 [wordlim: 200] Crawled: today;     `codex --model gpt-6.2-sol` ... GPT-6.2 Astra and GPT-6.2 Sol are now available in ChatGPT Work and Codex on supported plans and clients. ... The OpenAI API and Codex authenticated with your own API key aren’t affected by the GPT-5.4 retirement.
    
    6.2 Astra
    
    Choose GPT-6.2 Astra for demanding projects, such as diagnosing a bug across services or turning research into a presentation that follows your template.
    
    codex -m gpt-6.2-astra
    
    Copy command
    
    ChatGPT Work
    
    Web, mobile, and desktop
    
    Codex
    
    App and CLI
    
    Default context in Codex
    
    272,000 tokens
    
    6.2 Sol
    
    Use GPT-6.2 Sol for everyday tasks like editing code, drafting a report, or turning project notes into a plan.
    
    codex -m gpt-6.2-sol
    
    Copy command
    
    ChatGPT Work
    
    Web, mobile, and desktop
    
    Codex
    
    App and CLI
    
    Default context in Codex
    
    500,000 tokens
    
    Astra
    
    Combine reasoning and computer use to tackle complex work across code, apps, and research.
    
    codex -m gpt-6-astra
    
    Copy command
    
    Capability
    
    Speed
    
    ChatGPT desktop app
    
    ChatGPT web
    
    Codex CLI
    
    Codex IDE extension
    
    ChatGPT Credits
    
    
    6.1 Sol
    
    Performance close to GPT-6 Astra for complex work at a lower cost. Consider it for repeated, long-running work across code, apps, and documents.
    
    codex -m gpt-6.1-sol
    
    Copy command
    
    Capability
    
    Speed
    
    Work and Codex in the desktop app
    
    ChatGPT Work on the web
    
    Codex CLI
    
    Codex IDE extension
    
    ChatGPT Credits
    
    API Access
    
    ## Choosing a model
    
    Start with a task you know well and compare the result with what you need. Consider how long you’re willing to wait, how much usage the task requires, and the level of detail or judgment involved. You can change models as your work changes.
    
    ### Where each model shines
    
      * GPT-6.2 Astra or GPT-6 Astra, for the hardest end-to-end work. Choose an available Astra model for tasks that need sustained reasoning across code, apps, or research. Give it the sources, constraints, and checks that define a useful result.
      * GPT-6.2 Sol, for everyday work. Choose Sol to edit code, draft documents, or turn several sources into a plan.
      * GPT-6.1 Terra, for everyday writing and visual work. Choose Terra for clear communication, interpreting images and screens, and completing tasks across supported apps. Try it for drafting proposals, reviewing visual materials, or filling out forms from source documents.
      * Luna, for clear, repeatable tasks. Choose Luna for specific, high-volume tasks when you know what a good result looks like, such as extraction, classification, transformation, and structured summaries.
    --------------------------------------------------------------------------------
    ChatGPT & Codex changelog | ChatGPT Learn (https://learn.chatgpt.com/docs/changelog)
    citeturn2search2 [wordlim: 200] Crawled: today; GPT-6.2 Astra and GPT-6.2 Sol are now available in ChatGPT Work and Codex. ... In the CLI, use `/model` or start a task with `codex --model gpt-6-sol` or `codex --model gpt-6-luna`. ...     `$ npm install -g @openai/codex@0.153.3`
    
    # ChatGPT & Codex changelog
    
    Latest updates to ChatGPT and Codex
    
    All updates General ChatGPT desktop app Codex Mobile Codex CLI
    
    November 2026 October 2026 September 2026 August 2026 July 2026 June 2026 May 2026 April 2026 March 2026 February 2026 January 2026 December 2025 November 2025 October 2025 September 2025 August 2025 June 2025 May 2025
    
    ## November 2026
    
      * 2026-11-17
    
    ### GPT-6.2 Astra and GPT-6.2 Sol in Work and Codex
    
    GPT-6.2 Astra and GPT-6.2 Sol are now available in ChatGPT Work and Codex. Choose Astra for your hardest projects, such as implementing and testing a change across an existing codebase or turning a report into slides that follow your template. Choose Sol for everyday coding, writing, and multistep workflows. With GPT-6.2 Sol, you can complete more tasks on the same paid plan than with GPT-6.1 Sol.
    --------------------------------------------------------------------------------
    Changelog | OpenAI API (https://developers.openai.com/api/docs/changelog)
    citeturn2search3 [wordlim: 200] Crawled: today; Use GPT-6.2 Astra for demanding reasoning and coding tasks, or GPT-6.2 Sol for everyday coding and agent workflows. ... Approved organizations can use the OpenAI CSAM Image Classifier at no cost to help prioritize images for human review. ... Added markup-guided editing for supported GPT-Image-2.5 and GPT-Image-2.6 models.
    
    # Changelog
    
    Upcoming deprecations can be found on the deprecations page.
    
    ### November, 2026
    
    Nov 17
    
    Feature
    
    gpt-6.2-astra
    
    gpt-6.2-sol
    
    v1/responses
    
    v1/chat/completions
    
    Use GPT-6.2 Astra for demanding reasoning and coding tasks, or GPT-6.2 Sol for everyday coding and agent workflows. Both models support the Responses API and Chat Completions. Tool calling requires the Responses API.
    
    See the GPT-6.2 migration guide to update an existing integration.
    
    Nov 17
    
    Feature
    
    gpt-6.2-astra
    
    gpt-6.2-sol
    
    gpt-6.1-terra
    
    gpt-6-luna
    
    v1/responses
    
    Choose a two-hour or eight-hour background completion window in the Responses API public beta. Model token rates are 30% lower than Standard for the two-hour window and 40% lower for the eight-hour window.
    
    Nov 16
    
    Feature
    
    omni-moderation-csam-latest
    
    v1/moderations
    
    Approved organizations can use the OpenAI CSAM Image Classifier at no cost to help prioritize images for human review. The model returns separate signals for potential child sexual abuse material (CSAM) and potential AI-generated CSAM.
    
    Nov 5
    
    Feature
    
    gpt-6-cyber
    
    Use GPT-6 Cyber to investigate vulnerabilities, reproduce them in an authorized test environment, and verify fixes. See Daybreak setup before sending your first request.
    
    Nov 4
    
    Feature
    
    gpt-image-2.6-sunburst
    
    gpt-image-2.6-flare
    
    v1/images/generations
    
    v1/images/edits
    
    Released GPT-Image-2.6 Sunburst and GPT-Image-2.6 Flare for image generation and editing, with adjustable complexity, SVG generation, and animated GIF generation. Added markup-guided editing for supported GPT-Image-2.5 and GPT-Image-2.6 models.
    
    Nov 2
    
    Feature
    
    gpt-live-1
    
    v1/live/sessions
    
    Outbound SIP calling is now available to all organizations through the Live API. Connect GPT-Live to a phone number through your SIP provider. See the SIP guide.
    --------------------------------------------------------------------------------
    GPT-6.2 Astra Model | OpenAI API (https://developers.openai.com/api/docs/models/gpt-6.2-astra)
    citeturn2search4 [wordlim: 200] Crawled: today; Image: gpt-6.2-astra ... A frontier reasoning model for coding and professional work. ... GPT-6.2 Astra can build and revise software, use a computer, and work with documents.
    
    Models
    
    Image: gpt-6.2-astra
    
    GPT-6.2 Astra
    
    `gpt-6.2-astra`
    
    A frontier reasoning model for coding and professional work.
    
    A frontier reasoning model for coding and professional work.
    
    GPT-6.2 Astra can build and revise software, use a computer, and work with documents. It accepts text and images and responds with text. In the Responses API, use the Multi-agent beta to delegate work to subagents.
    
    The API rejects requests that set reasoning effort to `none`.
    
    The Responses API supports discounted background completion windows in public beta.
    
    1,050,000 context window
    
    922,000 max input tokens
    
    128,000 max output tokens
    
    May 18, 2026 knowledge cutoff
    
    Reasoning token support
    
    Pricing
    
    Prices are per 1M tokens. See the pricing page for processing tiers and billing details.
    
    Text tokens
    
    Per 1M tokens
    
    Input
    
    $8.00
    
    Cached input
    
    $0.40
    
    Cache writes
    
    $10.00
    
    Output
    
    $40.00
    
    The cache write price is for 30 minutes of retention. A 12-hour cache write costs $20 per million tokens for short-context requests at Standard pricing.
    --------------------------------------------------------------------------------
    Changelog | OpenAI API (https://developers.openai.com/api/docs/changelog?mw_entry=2026-09-29-99a06ddef837dc53)
    citeturn2search5 [wordlim: 200] Crawled: today; Use GPT-6.2 Astra for demanding reasoning and coding tasks, or GPT-6.2 Sol for everyday coding and agent workflows. ... Approved organizations can use the OpenAI CSAM Image Classifier at no cost to help prioritize images for human review. ... Released GPT-6 Astra, our most capable model, built for the hardest end-to-end work.
    --------------------------------------------------------------------------------
    Models | OpenAI API (https://developers.openai.com/api/docs/models)
    citeturn2search6 [wordlim: 200] Crawled: today; If you're not sure where to start, use GPT-6 Astra, our flagship model for complex reasoning and coding. ... All latest OpenAI models support text and image input, text output, multilingual capabilities, and vision. ... Sol GPT-6.1 Sol Near-Astra performance for complex work at a lower cost.
    --------------------------------------------------------------------------------
    Models | ChatGPT Learn (https://learn.chatgpt.com/docs/models?translationFallback=zh-Hans)
    citeturn2search7 [wordlim: 200] Crawled: today;     `codex --model gpt-6.2-sol` ... GPT-6.2 Astra and GPT-6.2 Sol are now available in ChatGPT Work and Codex on supported plans and clients. ... The OpenAI API and Codex authenticated with your own API key aren’t affected by the GPT-5.4 retirement.
    --------------------------------------------------------------------------------
    All models | OpenAI API (https://developers.openai.com/api/docs/models/all)
    citeturn2search8 [wordlim: 200] Crawled: today; GPT-6 Astra A reasoning model for complex coding and professional work.GPT-6.1 Sol Near-Astra performance for complex work at a lower cost.GPT-6 Sol Built to power complex coding and agentic workflows.GPT-5.6 Sol GPT-5.6 flagship model for complex professional work GPT-5.6 Terra GPT-5.6 model that balances intelligence and cost GPT-5.6 Luna GPT-5.6 model optimized for cost-sensitive workloads GPT-5.5 A new class of intelligence for coding and professional work.GPT-5.5 Pro Version of GPT-5.5 that produces smarter and more precise responses.GPT-5.4 A more affordable model for coding and professional work.GPT-5.4 Pro Version of GPT-5.4 that produces smarter and more precise responses.GPT-5.4 mini Our strongest mini model yet for coding, computer use, and subagents GPT-5.2 Previous flagship model for professional work with configurable reasoning effort GPT-5.2 Pro Previous pro model for professional work that produces smarter and more precise responses.omni-moderation Identify potentially harmful content in text and images
    --------------------------------------------------------------------------------
    Models | ChatGPT Learn (https://learn.chatgpt.com/docs/models?surface=web)
    citeturn2search9 [wordlim: 200] Crawled: 5 days ago;     `codex --model gpt-6.1-sol` ... 6 Sol Light Selections in this preview don't change your ChatGPT settings.6 Sol Light, 2 of 6. ... The six Power presets shown are Luna High, GPT-6 Sol Light (selected in this example), GPT-6 Sol Medium, Astra Light, Astra Medium, and Astra Extra High.
    --------------------------------------------------------------------------------
    Models | ChatGPT Learn (https://learn.chatgpt.com/docs/models?translationFallback=ja-JP)
    citeturn2search10 [wordlim: 200] Crawled: today;     `codex --model gpt-6.2-sol` ... GPT-6.2 Astra and GPT-6.2 Sol are now available in ChatGPT Work and Codex on supported plans and clients. ... GPT-5.5 remains available through the OpenAI API and in Codex sessions authenticated with an API key.
    --------------------------------------------------------------------------------
    Model selection | OpenAI API (https://developers.openai.com/api/docs/guides/model-selection?trk=article-ssr-frontend-pulse_x-social-details_comments-action_comment-text)
    citeturn2search11 [wordlim: 200] Crawled: today; Check the models available in ChatGPT or the API model catalog. ... Software engineering  | GPT-6.2 Astra, GPT-6.2 Sol, GPT-6.1 Terra, GPT-6 Luna
    
  • web__runweb__run

    input

    const r=await tools.web__run({open:[{ref_id:"turn2search0"}],response_length:"short"});text(r);
    

    result, 357 lines

    Script completed
    Wall time 2.2 seconds
    Output:
    GPT-6.2 | ChatGPT Learn (https://learn.chatgpt.com/docs/whats-new/gpt-62-astra-sol)
    citeturn3view0 [wordlim: 200] Crawled: today; Content type: text/html; Source: open({"ref_id":"turn2search0","lineno":null}); Total lines: 1021
    L25: ### Suggested
    L26: 
    L27: [Button: responses create][Button: reasoning_effort][Button: realtime][Button: prompt caching]
    L28: 
    L29: Primary navigation
    L30: 
    L31: [Button: API ][Button: ChatGPT ][Button: Docs ][Button: Use cases ][Button: Training ][Button: Resources ][Button: Resources ]
    L32: 
    L33: Theme Light [Input] Dark [Input] System [Input]
    L34: 
    L35: Search docs[Input: Start searching]
    L36: ### Suggested
    L37: 
    L38: [Button: responses create][Button: reasoning_effort][Button: realtime][Button: prompt caching]
    L39: 
    L40: [Button: Overview ][Button: Models ][Button: Agents ][Button: Tools ][Button: Audio & voice ][Button: Production ][Button: API reference ]
    L41: 
    L42: Docs Overview
    L43: 
    L44:   * 【2†Home】
    L45: 
    L46: ### Get started
    L47: 
    L48:   * 【14†Quickstart】
    L49:   * 【15†Using GPT-6.2】
    L50:   * 【16†Key concepts】
    L51:   * 【17†API credits for ChatGPT subscribers】
    L52: ### Core concepts
    L53: 
    L54:   * 【18†Responses API】
    L55:   * 【19†Decisions API】
    L56:   * 【20†Conversation state】
    L57:   * 【21†Background mode】
    L58:   * 【22†Streaming】
    L59:   * 【23†WebSocket mode】
    L60:   * 【24†Mid-turn steering】
    L61:   * 【25†Multi-agent】
    L62:   * 【26†Webhooks】
    L63:   * 【27†File inputs】
    L64:   * 【28†Compaction】
    L65:   * 【29†Counting tokens】
    L66: 
    L67: ### SDKs and CLI
    L68: 
    L69:   * 【30†OpenAI SDK】
    L70:   * 【31†OpenAI CLI】
    L71: 
    L72: ### Cloud providers
    L73: 
    L74:   * 【32†Amazon Bedrock】
    L75:   * 【33†Google Cloud】
    L76: ### Resources
    L77: 
    L78:   * 【34†Changelog】
    L79:   * 【35†Deprecations】
    L80:   * 【36†Supported countries】
    L81:   * 【37†OpenAI Crawlers】
    L82:   * 【38†Terms and policies†openai.com】
    L83: ### Legacy APIs
    L84: 
    L85:   * Agent Builder
    L86:     * 【39†Overview】
    L87:     * 【40†Migration guide】
    L88:     * 【41†Node reference】
    L89:     * 【42†Safety in building agents】
    L90:   * Evals
    L91:     * 【43†Getting started】
    L92:     * 【44†Working with evals】
    L93:     * 【45†Prompt optimizer】
    L94:     * 【46†External models】
    L95:     * 【47†Graders】
    L96:   * Fine-tuning
    L97:     * 【48†Optimization cycle】
    L98:     * 【49†Supervised fine-tuning】
    L99:     * 【50†Vision fine-tuning】
    L100:     * 【51†Direct preference optimization】
    L101:     * 【52†Best practices】
    L102:   * Assistants API
    L103:     * 【53†Migration guide】
    L733: ### Configuration
    L734: 
    L735:   * 【461†Windows app and CLI deployment】
    L752: ### ChatGPT EDU
    L753: 
    L754:   * 【475†Education plans】
    L755:   * 【476†Higher education setup】
    L756: 
    L757: ### ChatGPT Work
    L758: 
    L759:   * 【477†Overview】
    L760:   * 【478†Cloud security】
    L761:   * 【479†Local security】
    L762:   * 【480†Usage and cost】
    L763:   * 【481†Admin FAQ】
    L764: 
    L765: ### Collaboration and sharing
    L766: 
    L767:   * 【482†GPTs and sharing】
    L768: 
    L769: ### Plugins and connections
    L770: 
    L771:   * 【483†Plugin controls】
    L772:   * 【484†Import and sync plugins from GitHub】
    L773:   * 【485†Skill controls】
    L774:   * 【486†Migrate custom GPTs to plugins】
    L775: ### Usage and analytics
    L776: 
    L777:   * 【487†Workspace analytics】
    L778:   * 【488†Usage Insights】
    L779:   * 【489†Analytics API】
    L780: 
    L781: ### Security and compliance
    L782: 
    L783:   * 【490†Governance】
    L784:   * 【491†Agent security】
    L785:   * 【492†Safety and security in dots】
    L786:   * 【493†Prisma AIRS】
    L787:   * 【494†HIPAA configuration】
    L788:   * 【495†Compliance API and audit events】
    L789: 
    L790:   * 【5†Home】
    L791:   * 【496†All use cases】
    L792: 
    L793:   * 【7†Home】
    L794:   * 【497†Videos】
    L795:   * 【498†Showcase†developers.openai.com】
    L796:   * 【499†OpenAI Academy†openai.com】
    L797:   * 【500†Online trainings†academy.openai.com】
    L798: ### Community
    L799: 
    L800:   * 【501†Codex Ambassadors†developers.openai.com】
    L801:   * 【502†Codex for Students†developers.openai.com】
    L802:   * 【503†Codex for Open Source†developers.openai.com】
    L803:   * 【504†Events†luma.com】
    L804: 
    L805: ### Blog
    L806: 
    L807:   * 【505†Company blog†openai.com】
    L808:   * 【506†Developer blog†developers.openai.com】
    L809: 
    L810:   * 【5†Home】
    L811:   * 【496†All use cases】
    L812: ### Collections
    L813: 
    L814:   * 【507†Productivity & Collaboration】
    L815:   * 【508†Business Operations】
    L816:   * 【509†Data Science】
    L817:   * 【510†Web development】
    L818:   * 【511†Game development】
    L819:   * 【512†Native development】
    L820:   * 【513†Production systems】
    L821:   * 【514†Security】
    L822:   * 【515†Finance】
    L823:   * 【516†Sales】
    L824:   * 【517†Life Sciences】
    L825:   * 【518†Education】
    L826: 
    L827:   * 【7†Home】
    L828:   * 【497†Videos】
    L829:   * 【498†Showcase†developers.openai.com】
    L830:   * 【499†OpenAI Academy†openai.com】
    L831:   * 【500†Online trainings†academy.openai.com】
    L832: ### Community
    L833: 
    L834:   * 【501†Codex Ambassadors†developers.openai.com】
    L835:   * 【502†Codex for Students†developers.openai.com】
    L836:   * 【503†Codex for Open Source†developers.openai.com】
    L837:   * 【504†Events†luma.com】
    L838: 
    L839: ### Blog
    L840: 
    L841:   * 【505†Company blog†openai.com】
    L842:   * 【506†Developer blog†developers.openai.com】
    L843: 
    L844: 【519†Showcase】 [Button: Blog ][Button: Cookbook ][Button: Learn ][Button: Community ]
    L845: 
    L846: Docs Showcase
    L847: 
    L848:   * 【520†All posts】
    L849: ### Recent
    L850: 
    L851:   * 【521†Securing persistent agents like OpenClaw at OpenAI】
    L852:   * 【522†Pushing voice agents further: inside four GPT-Live builds】
    L853:   * 【523†Building a retro handheld game with Codex and ModRetro Chromatic】
    L854:   * 【524†Designing failover across OpenAI model providers】
    L855:   * 【525†Maintaining software with the GPT-6 family】
    L856: 
    L857: ### Topics
    L858: 
    L859:   * 【526†General】
    L860:   * 【527†API】
    L861:   * 【528†Apps SDK】
    L862:   * 【529†Audio】
    L863:   * 【530†Codex】
    L864:   * 【531†Life sciences】
    L865: 
    L866:   * 【532†Home】
    L867: ### Topics
    L868: 
    L869:   * 【533†Sign-in with ChatGPT】
    L870:   * 【534†Agents】
    L871:   * 【535†Evals】
    L872:   * 【536†Multimodal】
    L873:   * 【537†Text】
    L874:   * 【538†Guardrails】
    L875:   * 【539†Optimization】
    L876:   * 【540†ChatGPT】
    L877:   * 【541†Codex】
    L878:   * 【542†gpt-oss】
    L879: 
    L880: ### Contribute
    L881: 
    L882:   * 【543†Cookbook on GitHub†github.com】
    L883: 
    L884:   * 【8†Home】
    L885:   * 【544†OpenAI Developers plugin】
    L886:   * 【545†Docs MCP】
    L887: 
    L888: ### Categories
    L889: 
    L890:   * 【546†Demo apps】
    L891:   * 【547†Videos】
    L892: ### Topics
    L893: 
    L894:   * 【548†Agents】
    L895:   * 【549†Audio & Voice】
    L896:   * 【550†Computer Use】
    L897:   * 【551†Codex】
    L898:   * 【552†Evals】
    L899:   * 【553†gpt-oss】
    L900:   * 【554†Fine-tuning】
    L901:   * 【555†Image generation】
    L902:   * 【556†Scaling】
    L903:   * 【557†Tools】
    L904:   * 【558†Video generation】
    L905: 
    L906:   * 【559†Community】
    L907: 
    L908: ### Programs
    L909: 
    L910:   * 【560†Codex Ambassadors】
    L911:     * 【561†Meet the ambassadors】
    L912:   * 【562†Codex for Students】
    L913:   * 【563†Student Collective†openai.com】
    L914:   * 【564†Codex for Open Source】
    L915:   * 【565†OpenAI for Startups†openai.com】
    L916: ### Spaces
    L917: 
    L918:   * 【504†Events†luma.com】
    L919:   * 【566†Developer Forum†community.openai.com】
    L920:   * 【567†Discord†discord.com】
    L921:   * 【568†Reddit†www.reddit.com】
    L922:   * 【569†X†x.com】
    L923: 
    L924: 【570†API Dashboard†platform.openai.com】
    L925: 
    L926: 【571†Try ChatGPT†chatgpt.com】
    L927: 
    L928:   * 【4†Home】
    L929: 
    L930: ### Get started
    L931: 
    L932:   * 【309†Quickstart】
    L933:   * 【310†Use ChatGPT】
    L934:   * 【311†Get started with Work】
    L935:   * 【312†Meet dots】
    L936:   * 【313†Import from another agent】
    L937: ### Foundations
    L938: 
    L939:   * 【314†Prompting】
    L940:   * 【315†Working with agents】
    L941:   * 【316†Model selection】
    L942:   * 【317†Personalize ChatGPT】
    L943:   * 【318†Skills & Plugins】
    L944:   * 【319†Permissions】
    L945: 
    L946: ### Explore
    L947: 
    L948:   * 【320†What's new】
    L949:   * 【321†Models】
    L950:   * 【322†Pricing】
    L951:   * 【323†Glossary】
    L952: 
    L953: ### Available on
    L954: 
    L955:   * 【324†ChatGPT desktop app】
    L956:   * 【325†ChatGPT mobile app】
    L957:   * 【326†ChatGPT on the web】
    L958:   * 【327†Codex CLI】
    L959:   * 【328†Codex IDE extension】
    L960:   * 【329†Codex Cloud】
    L961: ### Releases
    L962: 
    L963:   * 【330†Changelog】
    L964:   * 【331†Feature Maturity】
    L965:   * 【332†Open Source】
    L966: 
    L967: 【572†Back to What's new】
    L968: 
    L969: 【573†Image】
    L970: 
    L971: November 17, 2026
    L972: 
    L973: # GPT-6.2
    L974: 
    L975: GPT-6.2 Astra and Sol launch in ChatGPT Work, Codex, and the API. Choose Astra for your hardest projects or Sol for everyday coding and writing.
    L976: 
    L977: ## What’s new
    L978: ### ChatGPT Work
    L979: 
    L980: Use Astra or Sol in Work on the web, desktop, or mobile with Plus, Pro, Business, Enterprise, or Edu. Enterprise and Edu admins must first 【574†enable the models in workspace settings】.
    L981: 
    L982: On supported consumer plans, new Work chats default to Sol at Light reasoning effort. Enterprise workspaces with model access default to Sol at Medium. If you see the Power slider, open Advanced to 【321†choose a model and reasoning effort】.
    L983: ### Codex
    L984: 
    L985: In the Codex app and CLI, choose Astra to trace a bug across services, fix it, and test the affected workflows. Choose Sol to review a pull request or revise code in response to feedback. Use the model picker in the app or `/model` in the CLI.
    L986: Codex and Work share usage. Pro $200 subscribers can use Ultrafast with Sol. Pro $500 subscribers can use it with either model, as can eligible Enterprise and Edu workspaces once an owner enables Ultrafast. See 【575†speed options】 and 【576†plan and client availability】.
    L987: ### Chat
    L988: 
    L989: To use GPT-6.2 Astra or Sol in ChatGPT, open Work. For the models you can select in Chat, see 【577†models in ChatGPT†help.openai.com】.
    L990: ### Responses API
    L991: 
    L992: Use `gpt-6.2-astra` or `gpt-6.2-sol` in the Responses API. Both models support Ultrafast at launch.
    L993: 
    L994: In standard, single-agent mode, you can change reasoning effort or switch from plain text to structured output between replies. Append a `configuration_update` item before the next user message and keep the request-level settings unchanged. 【167†Keeping earlier input intact】 helps preserve reasoning across turns and keeps that input eligible for prompt caching.
    L995: If your project uses data residency, both models support US and EU regions for Standard, Fast, and Batch processing. Ultrafast supports US data residency only. See the 【578†GPT-6.2 migration guide】 and 【579†API pricing by service tier】.
    L996: ## Choose your model
    L997: 
    L998: ### GPT-6.2 Astra
    L999: 
    L1000: Astra is the stronger choice for complex reasoning. It can work through a research question across multiple sources and check its conclusions against the data.
    L1001: #### Example use cases
    L1002: 
    L1003:   * Diagnose a bug across services. Give Astra the failing workflow and ask it to trace the cause, implement a fix, and run the affected tests.
    L1004:   * Build a deck from a report. Give Astra the report and your slide template. Ask it to present the findings in your team’s layout and writing style.
    L1005:   * Investigate a research question. Give Astra the sources and ask it to compare their findings and identify where the evidence disagrees.
    L1006: ### GPT-6.2 Sol
    L1007: 
    L1008: Sol handles everyday coding, writing, and work that takes several steps. Compared with GPT-6.1 Sol, it lets you complete more tasks on the same paid plan.
    L1009: #### Example use cases
    L1010: 
    L1011:   * Review a pull request. Ask Sol to walk through a change and flag issues for the author to address.
    L1012:   * Revise code after review. Give Sol the reviewer’s feedback and ask it to update the implementation and run the relevant tests.
    L1013:   * Turn meeting notes into a project brief. Ask Sol to pull together decisions, next steps, and open questions from your notes.
    L1014: 
    L1015: Explore the benchmarks and safety evaluations in the 【580†release announcement†openai.com】.
    L1016: 
    L1017: Ask AI
    L1018: ## Docs agent
    L1019: 
    L1020: Loading docs agent...
    
  • exec_commandls -la /opt/ms-playwright && ls -la /opt/node/lib/node_modules && python -c 'import PIL, numpy; print(PIL.__version__, numpy.__version__)' && ls -la .codex (+1 more)

    command

    ls -la /opt/ms-playwright && ls -la /opt/node/lib/node_modules && python -c 'import PIL, numpy; print(PIL.__version__, numpy.__version__)' && ls -la .codex
    
    fc-list | rg 'Nimbus|P052|DejaVuSansMono' | head -30 && env | cut -d= -f1 | rg 'MODEL|CODEX|ARTIFACT' 

    result, 4 lines

    Script completed
    Wall time 0.4 seconds
    Output:
    {"status":"fulfilled","value":{"chunk_id":"7fe3bf","wall_time_seconds":0.000003711,"exit_code":1,"original_token_count":268,"output":"total 40\ndrwx------ 12 runner runner  240 Nov 18 00:42 .\ndrwxr-xr-x  6 runner runner  120 Nov 18 00:42 ..\ndrwxr-xr-x  3 runner runner 4096 Apr  5  2026 chromium-1200\ndrwxr-xr-x  3 runner runner 4096 Sep  8 17:12 chromium-1208\ndrwx------  3 runner runner 4096 Oct 23 16:31 chromium-1248\ndrwxr-xr-x  3 runner runner 4096 Apr  5  2026 chromium_headless_shell-1200\ndrwxr-xr-x  3 runner runner 4096 Sep  8 17:12 chromium_headless_shell-1208\ndrwx------  3 runner runner 4096 Oct 23 16:31 chromium_headless_shell-1248\ndrwxr-xr-x  2 runner runner 4096 Feb 18  2026 ffmpeg-1011\ndrwx------  2 runner runner 4096 Oct 23 16:31 ffmpeg-1013\ndrwx------  3 runner runner 4096 Nov  6 18:20 firefox-1555\ndrwx------  4 runner runner 4096 Nov  6 18:20 webkit-2370\ntotal 16\ndrwx------ 4 runner runner 4096 Nov 11 07:55 .\ndrwx------ 3 runner runner 4096 Nov 11 07:55 ..\ndrwxr-xr-x 4 runner runner 4096 Oct 28  2025 corepack\ndrwxr-xr-x 7 runner runner 4096 Oct 28  2025 npm\nTraceback (most recent call last):\n  File \"<string>\", line 1, in <module>\nModuleNotFoundError: No module named 'numpy'\n"}}{"status":"fulfilled","value":{"chunk_id":"707fd9","wall_time_seconds":0.231655747,"exit_code":0,"original_token_count":709,"output":"/usr/share/fonts/truetype/dejavu/DejaVuSansMono.ttf: DejaVu Sans Mono:style=Book\n/usr/share/fonts/opentype/urw-base35/NimbusMonoPS-Italic.otf: Nimbus Mono PS:style=Italic\n/usr/share/fonts/opentype/urw-base35/P052-Italic.otf: P052:style=Italic\n/usr/share/fonts/type1/gsfonts/n021024l.pfb: Nimbus Roman No9 L:style=Medium Italic\n/usr/share/fonts/opentype/urw-base35/NimbusSans-Regular.otf: Nimbus Sans:style=Regular\n/usr/share/fonts/opentype/urw-base35/NimbusSansNarrow-Oblique.otf: Nimbus Sans Narrow:style=Oblique\n/usr/share/fonts/type1/urw-base35/NimbusSans-Regular.t1: Nimbus Sans:style=Regular\n/usr/share/fonts/type1/gsfonts/n021004l.pfb: Nimbus Roman No9 L:style=Medium\n/usr/share/fonts/opentype/urw-base35/NimbusMonoPS-BoldItalic.otf: Nimbus Mono PS:style=Bold Italic\n/usr/share/fonts/type1/gsfonts/n022003l.pfb: Nimbus Mono L:style=Regular\n/usr/share/fonts/opentype/urw-base35/NimbusSansNarrow-BoldOblique.otf: Nimbus Sans Narrow:style=Bold Oblique\n/usr/share/fonts/type1/urw-base35/NimbusSansNarrow-BoldOblique.t1: Nimbus Sans Narrow:style=Bold Oblique\n/usr/share/fonts/opentype/urw-base35/NimbusSans-Bold.otf: Nimbus Sans:style=Bold\n/usr/share/fonts/type1/urw-base35/NimbusMonoPS-Italic.t1: Nimbus Mono PS:style=Italic\n/usr/share/fonts/type1/urw-base35/NimbusRoman-BoldItalic.t1: Nimbus Roman:style=Bold Italic\n/usr/share/fonts/opentype/urw-base35/NimbusSans-BoldItalic.otf: Nimbus Sans:style=Bold Italic\n/usr/share/fonts/type1/gsfonts/n021023l.pfb: Nimbus Roman No9 L:style=Regular Italic\n/usr/share/fonts/type1/gsfonts/n019063l.pfb: Nimbus Sans L:style=Regular Condensed Italic\n/usr/share/fonts/opentype/urw-base35/NimbusRoman-Regular.otf: Nimbus Roman:style=Regular\n/usr/share/fonts/opentype/urw-base35/NimbusMonoPS-Bold.otf: Nimbus Mono PS:style=Bold\n/usr/share/fonts/opentype/urw-base35/NimbusRoman-Bold.otf: Nimbus Roman:style=Bold\n/usr/share/fonts/type1/urw-base35/NimbusSansNarrow-Bold.t1: Nimbus Sans Narrow:style=Bold\n/usr/share/fonts/opentype/urw-base35/NimbusRoman-BoldItalic.otf: Nimbus Roman:style=Bold Italic\n/usr/share/fonts/type1/gsfonts/n019064l.pfb: Nimbus Sans L:style=Bold Condensed Italic\n/usr/share/fonts/type1/urw-base35/NimbusRoman-Italic.t1: Nimbus Roman:style=Italic\n/usr/share/fonts/opentype/urw-base35/NimbusSans-Italic.otf: Nimbus Sans:style=Italic\n/usr/share/fonts/type1/gsfonts/n022023l.pfb: Nimbus Mono L:style=Regular Oblique\n/usr/share/fonts/truetype/dejavu/DejaVuSansMono-Bold.ttf: DejaVu Sans Mono:style=Bold\n/usr/share/fonts/type1/urw-base35/NimbusRoman-Regular.t1: Nimbus Roman:style=Regular\n/usr/share/fonts/opentype/urw-base35/NimbusMonoPS-Regular.otf: Nimbus Mono PS:style=Regular\nCODEX_MANAGED_PACKAGE_ROOT\nCODEX_VERSION\nCODEX_MANAGED_BY_NPM\nCODEX_SESSION_ID\nCODEX_HOME\nCODEX_THREAD_ID\nCODEX_APPLY_PATCH_PRESERVE_LINE_ENDINGS\nCODEX_CI\nCODEX_TOOL_CALL_ID\nCODEX_SANDBOX_NETWORK_DISABLED\n"}}