Fast, adaptive and willing to start

ChatGPT is unusually good at turning an incomplete request into momentum. It can ask questions, propose a structure and make a first version quickly.

• Broad multimodal capability
• Strong iterative conversation
• Useful from blank page to final artifact

The work may change shape

Start in Chat to clarify, move into Work for the finished artifact, or hand the software outcome to Codex without rebuilding the whole brief.

• The brief is still ambiguous
• Several output types may be needed
• The work benefits from shared context

Research and analysis

Give Work the decision or finished state, not just a question. It can gather current evidence, inspect files and keep a longer analytical process moving.

• Use connected apps and uploaded files
• Track progress instead of waiting blindly
• Require source and uncertainty reporting

Sites public beta

Sites extends the output from a static answer into dashboards, trackers, launch calendars, prototypes, internal portals and reports. The help article describes the public beta as available “for ChatGPT workspaces, Plus, and Pro accounts,” and says Sites “is not available on Free or Go, or in the EEA, Switzerland, or the United Kingdom at launch.” Read “workspaces” as the Business and Enterprise route rather than as a separate plan name.

• Not available in the EEA, Switzerland or the United Kingdom
• The help article carries no date, so read the eligibility as current state rather than a dated change, and confirm it in your own plan picker
• Treat live information and access as product requirements

Progress and steering

You can follow the work, answer questions and change direction without restarting the task or losing the developing context.

• Set review points early
• Redirect when assumptions change
• Do not mistake visible activity for quality

Local files, apps and browser

On desktop, Work can use local files and applications alongside a built-in browser. Computer Use lets it click, type and move files across your apps in the background. Since July 16 a unified Recents list merges Chat and Work, and Projects are available on desktop.

• Cloud Work conversations sync across web, mobile and desktop
• Local conversations stay local, so plan for where the record lives
• Grant only the context the task needs and review external transmissions

Auto-review of important actions

Auto-review uses OpenAI’s most advanced models to review consequential actions involving connected tools and APIs before they happen. OpenAI reports that in its own adversarial red teaming the feature blocked 100% of attempts to extract protected data.

• Vendor red-team result, not an independent audit
• A control, not a guarantee
• Compliance API gives oversight at scale
• Admins set spend and access limits

GPT-6 Astra New flagship · generally available

Launched September 3, 2026 as a new family, not a point release and not a rename. OpenAI’s model guidance calls it “our most intelligent model yet, with state-of-the-art performance in computer use, browsing, software engineering, science, and professional work,” and claims “Astra achieves stronger results while using substantially fewer output tokens, delivering a lower estimated API cost per task than earlier models.” Both are vendor claims. Astra is not a cost ladder: there is no Astra equivalent of Terra or Luna. It is no longer a single model either, because GPT-6 Astra Law shipped on September 17, 2026. On September 9, 2026 OpenAI published that Astra is “now available in ChatGPT Work, Codex, and the API,” and the developer models index now presents it as the flagship, “Our most capable model, built for the hardest end-to-end work.” The help center still warns that “Astra can use your allowance faster than GPT-5.6 Sol.”

• API id gpt-6-astra; 1,050,000 token context, 128,000 max output, knowledge cutoff April 30, 2026
• Standard short context $10.00 input / $1.00 cached / $50.00 output per 1M tokens
• Text in and out; images input only, with no image output and no audio
• Reasoning effort low, medium, high, xhigh and max; no fine-tuning
• Available since September 9, 2026: included with Plus and Pro plans and with Business Standard and Premium seats, Enterprise off by default at launch, Free, Go and Edu still unnamed

GPT-5.6 Sol GPT-5.6 flagship

Astra did not retire it. Sol is not deprecated, not legacy and unchanged in status, and OpenAI still describes it as delivering state-of-the-art results across coding, knowledge work, cybersecurity and science. Read its price off the Standard table: at short context that is $4.00 input and $20.00 output per 1M tokens. The $2.00 and $10.00 sitting one column over are the Batch and Flex figures, and they look exactly like a plausible list price. The promotional footnote is unchanged and still live: “GPT-5.6 Sol’s promotional pricing is available at least through November 21, 2026.”

• Standard short context $4.00 input / $0.40 cached / $20.00 output; long context $8.00 / $0.80 / $30.00; cache writes $5.00 short and $10.00 long
• Batch and Flex halve that; Fast mode doubles it
• Promotional: guaranteed only through November 21, 2026, so diarize that date
• Sol Pro available to Pro and Enterprise in Chat; Business Standard’s 15 Pro messages a month are shared across GPT-6 Pro and GPT-5.6 Sol Pro
• Still the safe default for volume work, and the place Astra traffic falls back to when an allowance runs out

GPT-5.6 Luna Most efficient

The fastest and most affordable tier, and the biggest mover on July 30, 2026, when OpenAI cut it by 80%. OpenAI reports it outperforms Claude Opus 4.8 on the Artificial Analysis Coding Agent Index at roughly a quarter of the estimated cost, a vendor comparison published on July 9. Its Standard rate is also exactly double its Batch rate, so check the column heading here too.

• Standard short context $0.20 input / $0.02 cached / $1.20 output; long context $0.40 / $0.04 / $1.80; cache writes $0.25 short and $0.50 long
• Batch and Flex: $0.10 / $0.01 / $0.60 at short context; Fast mode doubles Standard
• Became the default model for Free and Go on August 6, 2026, with unlimited text chats and a Think button
• Vendor benchmark, so verify on your own tasks

Prompt caching changes

GPT-5.6 introduces explicit cache breakpoints and a 30-minute minimum cache life. Note the billing change: cache writes cost 1.25x the uncached input rate. On September 3, 2026 the Responses API also gained the ability to change reasoning effort mid-conversation while preserving cached prompt prefixes, which removes a reason to throw a cache away.

• Cache reads keep the 90% discount
• Writes are no longer free, and every figure is now published: Astra $12.50 short and $25.00 long, Sol $5.00 and $10.00, Terra $2.50 and $5.00, Luna $0.25 and $0.50, each exactly 1.25x the uncached input rate
• Model your caching before you scale

Astra is generally available, and the entitlement is narrower than “generally available” sounds

Astra is available in ChatGPT Work, Codex and the API as of September 9, 2026. But “available” is not “everyone.” Work and Codex usage is included with Plus and Pro plans and Business Standard and Premium seats. Enterprise was off by default at launch and needs an administrator, and nothing dated since confirms that default has flipped, so an Enterprise buyer should verify rather than assume. Free, Go and Edu are named nowhere and stay unconfirmed. GPT-6 Pro in Chat excludes Plus. The practical failure mode is assuming a plan has it when the entitlement table says otherwise. An admin can settle it in a minute using the model access testing control added on September 11.

Four migration blockers stand in front of your first Astra test

OpenAI flags all four itself, and they matter more to a buyer than the benchmark table. none reasoning effort is not supported. Custom temperature, top_p and top_logprobs must be removed. Tool calling requires the Responses API rather than Chat Completions. And service_tier: fast is not supported with EU data residency. Any one of them can turn “try the new model” into a sprint. One honesty note on the evidence: all four come from the September 3 changelog entry, and the September 18 reading of the GPT-6 Astra model reference did not surface them again. They have not been contradicted anywhere, but they have not been reconfirmed either, so treat each as a thing to test on a single call rather than a settled fact.

Plausible citations

A link can be real while failing to support the claim placed beside it. Open the material sources.

Long-thread drift

Instructions and definitions can soften across long work. Restate the acceptance criteria at major transitions.

Memory boundaries

Review what is stored, which workspace is active and whether personal context belongs in managed work.

Regional and eligibility limits

Sites is unavailable in the EEA, Switzerland and the United Kingdom. Health is United States only, web and iOS, 18 and over, and is not available in Codex. Presence is limited to eligible enterprise customers. Confirm availability before you plan a rollout around a feature.

Image truth

Generated visuals may contain inaccurate text, products, interfaces or implied evidence.

September 25: new custom GPTs stop being creatable, on every plan

This is the largest migration in the window, and its first deadline is seven days out. Announced September 11, 2026: OpenAI is “planning to retire custom GPTs across ChatGPT plans and provide a migration path to plugins,” and “the transition affects all ChatGPT plans.” September 25, 2026 is when “creation of new custom GPTs ends.” December 11, 2026 is the “scheduled retirement” when “custom GPTs stop running.” Between those dates a migration experience moves a GPT’s instructions into a plugin as a skill. Here is the part that turns a migration into a rebuild: GPT custom actions do not transfer through the migration workflow. If your custom GPTs are prompt wrappers, this is an afternoon. If any of them calls an external API, that integration has to be written again as a plugin, and you have until December 11. The enterprise timeline is published as subject to change on an undated help center FAQ and a rolling release-notes entry, so read both yourself before you set an internal date.

September 24 is six days away, still has no successor, and there is an export window you have not used

The Sora 2 model line and the Videos API both go dark on September 24, 2026. The replacement column on the deprecations page was still empty when it was read on September 18, 2026, as it was on September 4, September 3, August 19 and August 3. Three separate OpenAI pages agree the date is live: the deprecations table, a banner on the video generation guide stating the Sora 2 models and Videos API “are deprecated and will shut down on September 24, 2026,” and a help article stating “the Sora API will be discontinued on September 24, 2026” and naming no successor. The Sora models are already absent from the models index. Now the part that is easiest to miss, and a buyer with content in Sora needs it this week: the Sora web and app experiences “were discontinued on April 26, 2026,” and OpenAI has published an export window at sora.chatgpt.com/sunset, after which “we will permanently delete any data associated with your use of Sora.” Export first, then plan the rebuild. The rebuild is not a version bump and you are six days from losing the source material as well as the API.

Both image waves point at a model that is already a generation behind

This is exactly the sort of thing that catches teams out, and it got worse on September 8. gpt-image-1 goes on October 23, 2026. Then gpt-image-1-mini, gpt-image-1.5 and chatgpt-image-latest go on December 1, 2026, announced June 2, 2026. A team that migrates in October and closes the ticket will be back in it five and a half weeks later. The new problem is the destination. Both waves still name gpt-image-2 for all four models, but OpenAI shipped ChatGPT Images 2.5 on September 8, 2026, with GPT-Image-2.5 Flare as the API default and GPT-Image-2.5 Sunburst for premium workflows. gpt-image-2 is now one generation behind, and it is priced well below the new pair: $2.50 input, $0.625 cached and $15.00 output per 1M against $8.00, $2.00 and $30.00 for the 2.5 models. Draw the consequence rather than leaving it implied. Anyone who follows the migration instruction exactly lands on a superseded model, and anyone who instead migrates to where OpenAI’s own product direction points lands on a bill roughly three times larger. Neither is wrong, but they are different decisions and the deprecations page does not tell you a choice exists.

Transcription has a 2027 deadline now

Announced August 26, 2026: on February 26, 2027, whisper-1, gpt-4o-transcribe, gpt-4o-mini-transcribe and gpt-4o-transcribe-diarize stop, replaced by gpt-live-transcribe or gpt-transcribe. A January 20, 2027 audio and realtime wave was announced on July 20, 2026. Neither is urgent; both belong on the roadmap now, because whisper-1 in particular is embedded in a great deal of quietly working infrastructure.

The November 30 wave is the one nobody has budgeted for

Three developer-platform pieces retire on November 30, 2026: the v1/prompts API and reusable prompt objects, the Evals dashboard and API, and Agent Builder. Two of the three replacements point away from OpenAI’s own surfaces: reusable prompts become “migrate to application code,” and Evals becomes “use Promptfoo alternative.” That is a different kind of deprecation from a model swap: it removes tooling teams have built process around.