Answer: Start with GPT-6.2 Sol for most tasks in ChatGPT Work or Codex, choose GPT-6.2 Astra for complex projects, use GPT-6 Luna for small repeatable tasks, and choose GPT-6.1 Terra when the task requires visual reasoning. The page also lists recommended tasks per model and reasoning effort, such as GPT-6 Luna · Low for small edits and field extraction, and GPT-6.2 Astra · Extra high for open-ended problems requiring trade-offs. Status: Confirmed Entry: tool/chatgpt/models/compare-models-and-reasoning-effort Revision: 2 Digest: sha256:83b47b270a1760af (changes whenever the content changes) Timeline: Created: 2026-11-13T13:24:32Z Last changed: 2026-11-18T13:34:15Z (GPT-6.2 Astra · Extra high tasks: — → GPT-6.2 Astra · Extra high) Last verified: 2026-11-20T13:40:28Z Next check: 2026-11-21T13:40:28Z Check interval: Daily Expected change: No fixed schedule Data: Field | Value | Source | As of Default model | GPT-6.2 Sol | [1] | 2026-11-20 Complex projects model | GPT-6.2 Astra | [1] | 2026-11-20 Small repeatable tasks model | GPT-6 Luna | [1] | 2026-11-20 Visual reasoning model | GPT-6.1 Terra | [1] | 2026-11-20 GPT-6 Luna · Low tasks | GPT-6 Luna · Low | [1] | 2026-11-20 GPT-6.2 Astra · Extra high tasks | GPT-6.2 Astra · Extra high | [1] | 2026-11-20 Sources: [1] OpenAI. "Model selection". https://learn.chatgpt.com/docs/model-selection.md Role: primary. Accessed: 2026-11-20T13:40:28Z Quote: "For most tasks in ChatGPT Work or Codex, start with **GPT-6.2 Sol**." (Supports: Default model) Quote: "Choose **GPT-6.2 Astra** for complex projects." (Supports: Complex projects model) Quote: "Use **GPT-6 Luna** for small, repeatable tasks, especially simple automations that run frequently." (Supports: Small repeatable tasks model) Quote: "Choose **GPT-6.1 Terra** when the task requires visual reasoning, such as navigating browser or desktop apps." (Supports: Visual reasoning model) Quote: "**GPT-6 Luna · Low**" (Supports: GPT-6 Luna · Low tasks) Quote: "**GPT-6.2 Astra · Extra high**" (Supports: GPT-6.2 Astra · Extra high tasks) Snapshot SHA-256: 9dcb65285297b85e3f0f16a198fd1f4bf82fd9547082d746f76c0fccf7eeff76 (captured 2026-11-18T13:34:12Z) Verification: Method: Automated extraction from [1] with deepseek-flash (JSON output, thinking disabled). Every value was checked by code to appear verbatim in a quoted passage of [1]. On this check the text of [1] was unchanged since 2026-11-18T13:34:15Z, so the values verified then were kept. Cross-check: not-available at 2026-11-20T13:40:28Z. No independent second source is configured for this entry Last successful check: 2026-11-20T13:40:28Z Last attempt: 2026-11-20T13:40:28Z Notes and limitations: Extracted automatically from a single official source (section "How to think about models and reasoning effort"). Related: https://latestn.com/tool/chatgpt/models/dalle-image-generation Can I still generate images with DALL·E on ChatGPT? https://latestn.com/tool/chatgpt/models/enable-gpt-6-2-astra-sol-enterprise-edu How do I enable GPT-6.2 Astra and Sol for my ChatGPT Enterprise or Edu workspace? https://latestn.com/tool/chatgpt/models/fedramp-api-methods-models Which API methods and models are available in ChatGPT FedRAMP? https://latestn.com/tool/chatgpt/models/choose-gpt-6-1-terra-use-cases Which ChatGPT model should I use for concise writing, computer-use tasks, or scoped code fixes? https://latestn.com/tool/chatgpt/models/choose-model-deep-research Can I choose a different model for deep research in ChatGPT? https://latestn.com/tool/chatgpt/models/choose-thinking-level How do I choose a thinking level for GPT-6 in ChatGPT? https://latestn.com/tool/ All entries in AI tools Change log: Time | Event | Change | Source 2026-11-18T13:34:15Z | changed | GPT-6.2 Astra · Extra high tasks: — → GPT-6.2 Astra · Extra high | [1] 2026-11-18T13:34:15Z | changed | GPT-6 Luna · Low tasks: — → GPT-6 Luna · Low | [1] 2026-11-18T13:34:15Z | changed | Visual reasoning model: — → GPT-6.1 Terra | [1] 2026-11-18T13:34:15Z | changed | Small repeatable tasks model: — → GPT-6 Luna | [1] 2026-11-18T13:34:15Z | changed | Complex projects model: — → GPT-6.2 Astra | [1] 2026-11-18T13:34:15Z | changed | Default model: — → GPT-6.2 Sol | [1] 2026-11-18T13:34:15Z | changed | Astra · Extra high use case: Demanding analysis and complex deliverables with exacting requirements. → — | [1] 2026-11-18T13:34:15Z | changed | Astra · Low use case: Concise writing and content adaptation that preserve facts and nuance. → — | [1] 2026-11-18T13:34:15Z | changed | GPT-6.1 Sol · Extra high use case: Polished deliverables, connected visual systems, and decisions built from conflicting evidence. → — | [1] 2026-11-18T13:34:15Z | changed | GPT-6.1 Sol · Medium use case: Complex technical work and coordinated deliverables you expect to revise. → — | [1] 2026-11-18T13:34:15Z | changed | Luna · Extra high use case: Finding current context across multiple apps, prioritizing work, and solving problems with clear constraints. → — | [1] 2026-11-18T13:34:15Z | changed | Luna · Low use case: Fine-grained edits, well-scoped problem-solving, and simple data extraction. → — | [1] 2026-11-18T13:34:15Z | changed | Most powerful model: Astra → — | [1] 2026-11-18T13:34:15Z | changed | Most efficient model: Luna → — | [1] 2026-11-13T13:24:32Z | added | Entry created | — Changes for this entry as a feed: https://latestn.com/tool/chatgpt/models/compare-models-and-reasoning-effort/latest Challenge this entry: Open challenges: None How: If a value here is wrong or out of date, send what is wrong and a link to evidence. Challenges are queued and checked against the evidence; they do not change the entry by themselves. If the entry changes as a result, the change log says so. Form: https://latestn.com/dispute/tool/chatgpt/models/compare-models-and-reasoning-effort POST: https://latestn.com/dispute/tool/chatgpt/models/compare-models-and-reasoning-effort with claim, evidence_url, and optionally field and contact (form-encoded or JSON) GET: https://latestn.com/dispute/tool/chatgpt/models/compare-models-and-reasoning-effort?claim={what is wrong}&evidence_url={link} How to cite: LatestN. (2026). How do I compare ChatGPT models and reasoning effort settings? Revision 2, last verified 2026-11-20T13:40:28Z. https://latestn.com/tool/chatgpt/models/compare-models-and-reasoning-effort (accessed 2026-11-21T10:03:34Z). @misc{LatestN_tool_chatgpt_models_compare_models_and_reasoning_effort_r2, title = {How do I compare ChatGPT models and reasoning effort settings?}, author = {{LatestN}}, year = {2026}, howpublished = {\url{https://latestn.com/tool/chatgpt/models/compare-models-and-reasoning-effort}}, note = {Revision 2; digest sha256:83b47b270a1760af; last verified 2026-11-20T13:40:28Z; accessed 2026-11-21T10:03:34Z} } Reuse: License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/) Attribution: Copying and adapting are allowed, including commercially, if you credit LatestN and link to the entry you used. Each entry ends with a ready-made citation. URL: https://latestn.com/tool/chatgpt/models/compare-models-and-reasoning-effort