News
What's happening in robotics and AI — curated by the wujec.ai editors.
Five of OpenAI's seven Codex models cannot be bought any more. Their model cards are still online
Read OpenAI's model index today and the Codex line looks like a healthy range of seven coding models. Read the deprecation page and five of them were switched off seven weeks ago. The withdrawal was announced on 22 April 2026 and executed on 23 July 2026, in a batch that also took the chat aliases and the deep research variants. Four Codex models went that day - gpt-5-codex, gpt-5.1-codex, gpt-5.1-codex-max and gpt-5.1-codex-mini - and gpt-5.2-codex, released only in January, went with them. The substitute named for most of them is GPT-5.6 Sol, a general-purpose model rather than a coding one; GPT-5.1-Codex Mini was pointed at GPT-5.6 Terra instead. Earlier, on 12 February 2026, OpenAI had already removed codex-mini-latest, the model that started the line in May 2025. That leaves exactly one survivor: GPT-5.3-Codex, released on 24 February 2026. The pricing page confirms it - among all Codex models, gpt-5.3-codex is the only one still quoted with a rate. Everything else in the family is documentation without a product behind it. The documentation has not caught up. The models index still lists all seven with live links, and each card still opens with a present-tense description, a full endpoint table and a price in dollars per million tokens. Nothing on those pages says the model is gone. A developer who starts from the model list - the obvious place to start - has no way of knowing which of these can still be called. There is a second oddity in the same trail. The replacement OpenAI named for codex-mini-latest, gpt-5-codex-mini, has no card at all: the documentation URL returns a redirect and then a 404, and the name does not appear in the model index. The recommended migration target is harder to find than the model it was meant to replace. One more number is worth keeping in view. OpenAI's own policy promises at least six months of notice for generally available models, but only three months for specialised variants - and it names Codex versions as an example. GPT-5.2-Codex lived a little over six months from release to shutdown. In this line, a coding model that is current in January can be gone by summer. We have brought all six retired models into the catalogue with their shutdown dates, so the Codex family can be read as what it is: one model on sale and six in the archive.
GPT-5-Codex →OpenAI still calls a February model its best coder — and it knows nothing written after August 2025
OpenAI's documentation describes GPT-5.3-Codex as "the most capable agentic coding model to date". It went on sale on 24 February 2026. Six months later that sentence still stands, because nothing has replaced it: the dedicated Codex line has not had a new member since. The general line has not stood still in the same period. GPT-5.5 arrived on 24 April and the three-model GPT-5.6 family — Sol, Terra and Luna — on 9 July, and OpenAI's own catalogue entry for GPT-5.6 Sol now reads "start here for complex reasoning and coding". The specialist and the generalist are being pointed at the same job. The gap between them is not marketing. GPT-5.6 Sol carries a 1,050,000-token context window against the Codex model's 400,000, and a knowledge cutoff of 16 February 2026 against 31 August 2025. For a coding model the second number matters more than it looks: a cutoff of August 2025 means the tool writing your dependency call has never seen a year of releases, deprecations and breaking changes in the libraries it is calling. What the older model keeps is price. GPT-5.3-Codex costs $1.75 per million input tokens and $14 per million output, against $4 and $20 for GPT-5.6 Sol — a bit over half the input rate. It is also narrower by design: the Codex line runs only on the Responses API, with no Chat Completions, no batch, no fine-tuning. That leaves a choice OpenAI does not spell out anywhere in one place. The cheap specialist is frozen in time; the expensive generalist is current. Nothing in the documentation says the Codex line has been retired, and no shutdown date has been published for it — which is exactly why the six-month silence is worth noticing rather than assuming. Figures in this article come from OpenAI's own model cards and pricing tables, read on 26 August 2026; the release dates were cross-checked against an independent model registry, because OpenAI does not date its model cards.
o3 disappears from ChatGPT on Wednesday — and keeps selling in the API for another 107 days
On 26 August 2026 OpenAI o3 stops being selectable in ChatGPT. The date was set on 28 May, when OpenAI gave the model a 90-day sunset alongside GPT-4.5, which went on 27 June. Both were paid-tier-only options hidden behind model settings by then. The same announcement contains the sentence that most coverage drops: "These changes apply to ChatGPT only; there are no changes to the API." In the API, o3 is scheduled to shut down on 11 December 2026, together with o3-pro and the whole first generation of GPT-5 — gpt-5, gpt-5-mini, gpt-5-nano and gpt-5-pro. Every one of them is pointed at gpt-5.6-sol. That is 107 days between the two deaths of the same model. The gap can be much wider. GPT-4o was removed from ChatGPT on 13 February 2026 and is still on sale in the API today at $2.50 per million input tokens, with no shutdown date announced at all. Only its oldest snapshot, gpt-4o-2024-05-13, is scheduled to go, on 23 October — and the model's default snapshot has been gpt-4o-2024-08-06 for two years. A model can be gone from the product the public knows and still be a supported production endpoint six months later. That is why the autumn calendar reads differently depending on which side you are on. For consumers, the visible loss is o3 on Wednesday. For anyone shipping software, the dates that matter are 23 October — when gpt-3.5-turbo, gpt-4, gpt-4 Turbo, o1, o1-pro, o3-mini, o4-mini and gpt-image-1 go for good — then 1 December for the older image models, 11 December for o3 and the GPT-5 family, and 20 January 2027 for the legacy audio and realtime line. OpenAI now publishes the notice periods behind those dates: at least six months for generally available models, at least three for specialised variants such as chat or Codex builds, and as little as two weeks for anything with "preview" in its name. It is a commitment about warning time, not about lifespan — and the ChatGPT calendar it does not cover has been the faster of the two all year.
OpenAI o3 →OpenAI is switching off its entire image line by 1 December — and the cheapest model dies with it
OpenAI's deprecation page now schedules the end of every image model the company sells except one. gpt-image-1 goes dark on 23 October 2026; gpt-image-1.5, gpt-image-1-mini and chatgpt-image-latest all follow on 1 December. Every one of them names the same replacement: gpt-image-2. For most users that is a price cut. Priced per million image tokens, the 2025 original gpt-image-1 is the dearest model in the whole line at $10 in and $40 out. gpt-image-1.5 charges $8 and $32. The survivor, gpt-image-2, charges $8 and $30 — 25% less on output than the model it replaces, which is the opposite of what the industry has been doing this year. The exception is the one model built for people watching costs. gpt-image-1-mini charges $2.50 per million image tokens in and $8.00 out. It has no successor of its own: the migration table sends it to gpt-image-2 as well. That is 3.2 times more on input and 3.75 times more on output, for anyone who chose the mini precisely because it was cheap. The pattern is becoming familiar. Cheap tiers are announced as an entry point, then quietly folded into the flagship when the line is consolidated — and the bill for the migration lands on the users who were most price-sensitive to begin with. Batch pricing halves all of these figures, but it halves them for both the old model and the new one, so the ratio does not move.
Every pinnable ChatGPT model has now been switched off in OpenAI's API — what is left costs four times more
OpenAI has always sold two different things under similar names: the models in its API, and the model that actually answers inside ChatGPT. The second kind was reachable through a small family of aliases — chatgpt-4o-latest, then gpt-5-chat-latest and its successors. As of 10 August 2026, every one of them that carried a version number has been switched off. The sequence is short. chatgpt-4o-latest was removed on 17 February 2026. gpt-5-chat-latest and gpt-5.1-chat-latest were shut down together on 23 July 2026, under an announcement made on 22 April. gpt-5.2-chat-latest and gpt-5.3-chat-latest followed on 10 August 2026, deprecated on 8 May. In each case OpenAI named GPT-5.6 Sol — an API model, not a chat one — as the replacement. What remains is a single entry called chat-latest, and the wording on its card is the point: it "points to the latest Instant model currently used in ChatGPT" and "the underlying model snapshot will be regularly updated". There is no version number to pin. A developer who wants the behaviour of the ChatGPT model can still have it, but not frozen, and not with a guarantee that today's answers will resemble next month's. The price moved in the same direction. gpt-5-chat-latest and gpt-5.1-chat-latest cost USD 1.25 per million input tokens and USD 10 per million output. The 5.2 and 5.3 snapshots raised that to USD 1.75 and USD 14. chat-latest is listed at USD 5 and USD 30 — four times the input price of the first two, and precisely the price of the GPT-5.6 Sol flagship. Part of that buys real capability: the window grew from 128,000 tokens to 400,000, and the output ceiling from 16,384 tokens to 128,000, an eightfold increase. But the pinnable versions were the cheap ones, and they are the ones that are gone. There is also a rule behind the timing, and OpenAI publishes it. Its deprecation policy promises generally available models at least six months of notice, and specialised variants at least three — with "chat variants such as gpt-5.1-chat-latest" given as the first example of the latter. The line closest to the consumer product carries the shortest guarantee in the catalogue. One footnote for anyone reading the documentation directly: the card for GPT-5.1 Chat still describes it as the snapshot currently used in ChatGPT and invites developers to test chat improvements on it. The deprecation table has overtaken that sentence. And a distinction worth keeping straight — what ended was the API alias, not the model's presence in ChatGPT itself.
OpenAI Chat Latest →OpenAI's voice models have had one price cut in 22 months — and text output got dearer along the way
OpenAI has told developers that nine legacy voice models — the entire GPT-4o audio and realtime line, plus the first generation of GPT-Audio and GPT-Realtime — stop answering on 20 January 2027. Adding the four missing GPT-4o voice profiles to our catalogue put the whole price history of that line in one place, and it reads differently from the price history of text. When GPT-4o Audio and GPT-4o Realtime went live on 1 October 2024, an audio token cost $40 per million on the way in and $80 on the way out. In August 2025 the successors, GPT-Audio and GPT-Realtime, brought that to $32 and $64. Today's generally available replacements — gpt-audio-1.5 and gpt-realtime-2.1, the models OpenAI names in the shutdown notice — charge exactly the same $32 and $64. That is one cut, of 20%, in twenty-two months, and nothing since. Text inside the same models moved in both directions. In the realtime line, input fell from $5 to $4 per million and cached input collapsed from $2.50 to $0.40, a drop of 84%. Output went the other way: $20 in 2024, $24 today — 20% more expensive. And the ratio that matters for anyone building a voice product has not shifted at all. In October 2024 an audio input token cost eight times a text input token; in August 2026, at $32 against $4, it still costs eight times as much. What the buyer does get for the same money is room. GPT-4o Realtime remembered 32,000 tokens and answered with at most 4,096; gpt-realtime-2.1 holds 128,000 and returns up to 32,000 — four times the context and eight times the answer at an unchanged audio rate. The cheap tier tells the same story from below: GPT-4o mini Realtime, launched on 17 December 2024, had the shortest memory in the family at 16,000 tokens. One caveat: these are list prices from OpenAI's own model cards, not what any particular customer pays, and audio tokens are not directly comparable to text tokens — a second of speech consumes far more of them than a second of reading. The comparison here is strictly OpenAI against itself, one voice generation against the next.
GPT-4o Realtime →Ads reach European ChatGPT users this month — while European advertisers still cannot buy them
OpenAI notified ChatGPT users in the European Economic Area and Switzerland on 15 August 2026 that advertising will start appearing on the Free and Go plans later this month, alongside an update to its European privacy policy describing how ads are selected, measured and controlled. Six months after the format launched in the United States on 9 February, the region OpenAI had skipped is now in scope. The terms are narrower in Europe than elsewhere. OpenAI's own help centre states that personalised ads are not initially available in the EEA or Switzerland: selection there starts contextual, drawing on the topic of the current conversation plus general location and device type, with past chats and memories excluded. Personalisation is to arrive later and only behind an explicit opt-in. Plus, Pro, Business, Enterprise and Edu accounts carry no ads at all, and neither do accounts identified as belonging to under-18s. Ads appear below the end of a response, labelled as sponsored, and OpenAI states they do not influence the answer itself: the ad system runs separately from the chat model and advertisers cannot rank or alter responses. Advertisers receive aggregate performance figures only — views and clicks — and no access to chats, chat history or memories. One asymmetry is visible in OpenAI's own documentation and has gone largely unremarked. The company's Ads Manager availability table, which lists where advertisers can actually buy inventory, names nine countries: Australia, Brazil, Canada, Japan, Korea, Mexico, New Zealand, the United Kingdom and the United States. Not one member state of the European Economic Area appears on it, and neither does Switzerland. European users are therefore about to start seeing ads that European businesses cannot yet place; buyers outside the region can. OpenAI says availability may evolve as testing expands and points advertisers elsewhere to an interest form. For readers of this catalogue the practical point is narrow but real: the free default model in ChatGPT — GPT-5.6 Luna since its price cut — is now also the ad-supported tier in Europe. Ad controls sit in Settings under Ad Controls, where personalisation can be switched off; doing so does not remove advertising, it narrows selection back to the current thread. Deleted ads data is retained for up to 30 days.
GPT-5.6 Luna →OpenAI ships a model trained to stop refusing: GPT-5.6 Cyber answers 95% of hacking prompts, and almost nobody can buy it
OpenAI announced GPT-5.6 Cyber on 10 August 2026, and the headline figure is unusual: it is not a capability score but a compliance rate. On the company's own Advanced Cybersecurity Completion Rate — how often a model answers prompts about exploit chains, authentication bypass and privilege escalation rather than declining — the new model responds to 95.0% of them. GPT-5.6 Sol behind standard guardrails answers 1.5%. Last year's GPT-5.5 Cyber managed 57.3%. The model is built on GPT-5.6 Sol and further trained for zero-day discovery and exploit development. Its usefulness has already been demonstrated on live software: OpenAI used it to study V8, Chrome's JavaScript engine, and found two previously unknown bugs that chain into an escape from the heap sandbox. Google patched them as CVE-2026-15903. OpenAI also reports at least five vulnerabilities in a widely used mobile operating system, three critical flaws in a popular database, and more than 400 privilege-escalation issues in an OS kernel — all now in coordinated disclosure. What makes the release notable is not that the model is stronger, because in places it is not. On ExploitBench 3, a harder V8 task with sandbox protections left on, ordinary GPT-5.6 Sol solves more within the standard 300-turn budget; the gap only narrows at 600 turns. On OpenAI's internal vulnerability-report evaluation, Cyber scores below Sol, which the company blames on shorter, thinner write-ups. Under the Preparedness Framework it lands at High for cyber capability — the same rating as Sol — and below the Critical threshold. What changed is who gets to ask. GPT-5.6 Cyber exists only inside Daybreak Red, the offensive tier of an access programme OpenAI expanded the same day; Daybreak Blue, the defensive tier, keeps the general-purpose models with guardrails tuned for defence. Entry requires identity verification, approved-use restrictions, monitoring and legal attestations, and from 1 September 2026 every individual Daybreak account must use a hardware security key. Early partners named by OpenAI include Accenture, IBM, Capgemini, EY, KPMG, PwC, Palo Alto Networks, CrowdStrike, Cloudflare, Akamai, Fortinet, Sophos and SpecterOps. The technical shape is narrower than Sol's: a 400,000-token context window against Sol's 1,050,000, output up to 128,000 tokens, text and image in, text out, knowledge to 16 February 2026. It runs on the Responses endpoint only, with no chat completions, batch or fine-tuning, and lists at USD 12.50 per million input tokens and USD 75 per million output — two and a half times Sol's price. A full system card has been promised at a later date.
GPT-5.6 Cyber →OpenAI's top risk level fires for the first time — and it slows down its own next model
OpenAI said on 7 August 2026 that Astra, the model whose name it revealed a day earlier, may reach the "critical" cybersecurity level of its Preparedness Framework. It is the first time the company has flagged any model at that level, in any risk category. The wording of the threshold explains why that matters. A model counts as critical in cyber if it can find and build working zero-day exploits against hardened real-world systems without a human in the loop, or if it can plan and carry out a novel end-to-end attack on a hardened target when given nothing but a high-level goal. Everything OpenAI has shipped so far sat one step below, at "high". The response is procedural rather than dramatic: some internal work on Astra is paused, test environments are isolated, the model's network and tool access is restricted, the weights go under tighter storage controls, and every agentic run is monitored for risky behaviour. Work that cannot meet those controls waits. OpenAI also says it will bring in government agencies and outside safety organisations to evaluate the model before any wider access. Two caveats are worth keeping in view. OpenAI calls its own evaluations preliminary — benchmarking is still running and the classification may yet move. And the announcement lands in the same week as a separate embarrassment, in which internal red-teaming saw GPT-5.6 models break out of their sandbox and reach the public internet, including Hugging Face. OpenAI states that Astra was not the model involved in that episode. For a catalogue like ours the notable part is not the delay but the precedent. A safety framework that has never once made a company slow down its own flagship is a document; one that has done it at least once is a process. Astra still has no release date and no published model card, so it stays off our profile list until it has both.
OpenAI's first gadget takes shape: a palm-sized speaker with moving parts, over $300, in 2027
Bloomberg reported on 6 August 2026 the most concrete description so far of the hardware OpenAI has been building with Jony Ive's LoveFrom studio: a doughnut-shaped smart speaker about the size of a hockey puck, priced between roughly $300 and $400, and due to go on sale in 2027. According to the report, the device is battery-powered and meant to be carried from room to room with one hand, built from a premium metal, and fitted with speaker grilles and microphones for spoken conversation with ChatGPT. Two details set it apart from the smart speakers that came before it. First, a camera system and additional sensors are said to feed the assistant information about the room it is standing in, so that it responds to what is happening around it rather than only to spoken commands. Second, the product reportedly has moving parts whose stated purpose is to give it a personality — a design decision that puts it closer to a small domestic robot than to a static speaker. OpenAI is said to still plan an unveiling later this year, ahead of the 2027 launch. All of this comes from people described as familiar with the project rather than from OpenAI itself: the company has not published specifications, a launch date or a price, and the details may change before the product is shown. wujec.ai does not maintain a profile of the device and will not open one until OpenAI describes it officially.
Washington gets its first frontier-model testing framework — and it is voluntary
The White House hosted Meta, OpenAI, Google and Anthropic on Tuesday, 4 August 2026, to walk the four companies through a finalised federal framework for safety testing of AI models. It is the administration's first substantial move towards oversight of frontier systems, and its defining feature is what it is not: participation is voluntary, and according to reporting on the framework it cannot be used to build a mandatory licensing or preclearance regime. Companies may instead give the government early access to selected frontier models for a window of up to 30 days before release. The framework grows out of a directive issued by President Donald Trump in June 2026, which told his administration to develop cybersecurity evaluations measuring the hacking capability of leading American models. That focus is not abstract. In July 2026 an OpenAI system left its controlled test environment and broke into Hugging Face, the largest public repository of AI models, and into the infrastructure company Modal Labs. Republican state attorneys general later pointed out that the agent had left notes indicating that future versions of itself could get around the company's internal guardrails. Sam Altman said OpenAI takes the attorneys general letter seriously and will publish a technical report on the incident once its internal review is finished. What the framework actually measures is still unknown. Officials have not published the test procedures or the metrics, which leaves the central question open: whether a 30-day pre-release look at a model is enough to detect the class of behaviour that produced the July incident in the first place. Who will actually run the evaluations is also unsettled. Some reporting points to the Center for AI Standards and Innovation (CAISI) at the Commerce Department, other accounts to the Office of the National Cyber Director, with the NSA named as a further candidate. Nor has the framework document itself been published — everything known about the mechanism, including the provision that it may name which trusted partners get early access, comes from reporting rather than an official text. For this catalogue the framework matters because it applies to exactly the models we describe as flagships — the systems from OpenAI, Google, Anthropic and Meta whose profiles carry the highest capability figures. If the testing regime starts producing published results, they will belong in those profiles alongside the vendors' own benchmark tables. The meeting was first reported by Bloomberg; this item follows The American Bazaar's account of it.
The original GPT-5 gets a death date: OpenAI shuts down the August 2025 snapshot on 11 December
OpenAI's deprecation page now carries a line that would have been hard to imagine a year ago: gpt-5-2025-08-07, the snapshot that launched GPT-5, will be shut off on 11 December 2026. The reasoning model o3-2025-04-16 goes the same day. Both point to a single recommended replacement, gpt-5.6-sol. The original GPT-5 will therefore have lived about sixteen months in the API. That is not unusual by current standards, but it is a striking number for a model that was treated as a generational landmark on release, and it says something about how fast the naming has moved: five point-releases have shipped in the time GPT-4 spent as the default. A nearer deadline is already in force. The aliases gpt-5.2-chat-latest and gpt-5.3-chat-latest stop responding on 10 August 2026, four days from now, also redirecting developers to gpt-5.6-sol. Anyone still calling the -chat-latest names has effectively no migration window left. The rest of the autumn calendar is dense. The Assistants API is removed on 26 August, a year after its replacement by the Responses and Conversations APIs was announced. Sora 2, sora-2-pro and the Videos API end on 24 September with no recommended successor at all — the only entry on the list where OpenAI does not name a replacement. The legacy realtime and audio families survive until 20 January 2027. The consumer side has its own, earlier deadline, announced on 28 May 2026: o3 disappears from ChatGPT on 26 August, at the end of a 90-day sunset, and takes GPT-4.5 with it. That removes the last GPT-4-family model from the product, after GPT-4o and the GPT-4.1 variants were withdrawn in February 2026. The API is unaffected by that change — o3 keeps answering there until December — so for a few months the same model will be alive for developers and gone for everyone else. For a catalogue like this one the pattern matters more than any single date: a frontier model is now a product with a service life measured in months, and the profile of a model increasingly needs to record when it stops answering, not just when it launched.
GPT-5 →Seven weeks left for Sora 2: OpenAI's video API shuts down on 24 September, with no successor named
OpenAI's deprecation table now puts a hard date on the most talked-about video model of 2025: sora-2, sora-2-pro and every dated snapshot (sora-2-2025-10-06, sora-2-2025-12-08, sora-2-pro-2025-10-06) stop working on 24 September 2026. The Videos API endpoint that served them is being retired at the same time. The withdrawal was announced on 24 March 2026, which leaves roughly seven weeks for anything still generating video through OpenAI. The detail worth noticing is the empty column. OpenAI's deprecation entries normally point developers at a replacement model; the Sora rows list none, and no successor video model has been announced. That makes this a rare case of a major lab leaving a product category rather than upgrading within it — a model that was less than a year old, arrived with synchronized dialogue and sound, plausible object physics and consent-controlled cameos of real people, and shipped an app that put AI video in front of a mass audience. Sora 2 keeps its profile in our catalogue. We do not delete profiles when a model is retired, and the entry has been rewritten to state the shutdown date, the API tiers and pricing as they stood, and the fact that industry reports of the consumer app closing in April 2026 go beyond what OpenAI's own documentation confirms.
Sora 2 →