OpenAI’s $200 Chromatic brings Codex games to cartridges
Editorial Team
••📖6 min read
OpenAI’s $200 Chromatic: Codex Edition adds AI-assisted development to an existing cartridge-game platform. The public bundle includes a programmable cartridge, but creation requires a computer, established tools and playtesting; the store now lists a November 19 restock.
The $200 Codex Edition includes a programmable cartridge; OpenAI’s archived listing gives November 19 as its expected restock date.
Games are created on a computer with Codex and existing development tools, then tested in an emulator or transferred to the handheld.
Cartridge writing requires a firmware update and Developer Mode activation, and can erase an existing game and its saves.
AI developer OpenAI has opened public sales of the ModRetro Chromatic: Codex Edition, a $200 retro handheld from gaming hardware maker ModRetro, bundled with a programmable cartridge for games made using OpenAI’s Codex coding agent. Previously a DevDay developer-conference exclusive, the collaboration is now a retail product, although OpenAI’s store listing already shows a restock notice.
The offer connects AI-assisted software creation to a physical game users can carry. It builds on two existing products: ModRetro’s Game Boy-compatible Chromatic handheld and OpenAI’s coding agent. Users develop on a computer through the Codex desktop app, test and revise the result, then write the game to a cartridge for play without the computer.
ModRetro’s quickstart shows a game running in Codex’s browser preview. The handheld-shaped display is an emulator interface. Credit: ModRetro. Source: ModRetro Support.
An existing handheld gets a new development workflow
OpenAI’s product listing describes the public edition as a new look for the earlier DevDay exclusive. It includes one programmable game cartridge and uses a transparent ABS plastic shell, a 2.56-inch backlit IPS display with 160 × 144 pixels, and three AA batteries. A ModRetro rechargeable power accessory is also compatible.
ModRetro, the retro-gaming hardware maker supplying the handheld, announced the original Chromatic on June 3, 2024 as a Game Boy cartridge-compatible device. That company announcement already described a route for independent developers to apply to have cartridge games manufactured and sold. Creating new games for the platform was part of its earlier proposition; Codex adds a conversational way to build them.
UC Berkeley and NVIDIA are releasing tools for AI agents to assemble and improve robot programs in simulation. A warehouse operator reports a useful production update, while unsuccessful fixes and benchmark results show why testing, speed and existing infrastructure still matter.
Editorial Team
OpenAI introduced the current Codex agent as a research preview in May 2025, describing software-engineering tasks such as writing features and fixing bugs. The Chromatic workflow applies that coding role to Game Boy Color-compatible games. The Codex Edition also plays Game Boy, Game Boy Color and Chromatic cartridges, giving buyers an existing game library alongside their own creations.
The public release went on sale November 9, according to a launch report that dates the conference introduction to September 29, 2026. OpenAI’s retained store page now says “Restock soon,” with an expected November 19 return. Shipping is limited to the United States, Canada, Australia, New Zealand, Japan, the United Kingdom and the European Union.
The $200 price needs a hardware comparison. A separate launch-day report puts a standard Chromatic at $199.99 and says a transparent model has sold for $149.99 since October. Those are handheld prices, rather than equivalent development bundles: the Codex Edition explicitly includes its programmable cartridge and, under ModRetro’s setup instructions, a Developer Mode activation code. The figures do not establish the full cost of computer equipment or Codex usage.
Lodge Podge open in GB Studio, where scenes can be edited before rebuilding the game with Codex. Credit: OpenAI. Source: OpenAI Developers.
From a prompt to a playable ROM
OpenAI’s developer guide starts in the Codex desktop app. Users install the ModRetro Chromatic plugin, select a local project folder and describe a game. The plugin provides tools for building artwork, maps and game logic, compiling a ROM—the game file—and opening it in a browser emulator.
GB Studio, the visual Game Boy game editor used in the workflow, remains available for direct editing. Users can open the project, move characters or change collision boundaries, save it, and ask Codex to rebuild. The plugin can also compile games written in C using GBDK-2020; assembly projects require a separately installed toolchain such as RGBDS.
The output stays within the existing Game Boy development ecosystem: a compatible game built with established tools. What Codex adds is a conversational interface for coordinating them: generating sprite sheets, changing blocked tiles, annotating a frozen game frame and running sequences of button presses for inspection.
Testing and deployment also need separating. A game can run in the browser before a user buys hardware. ModRetro’s quickstart guide says live testing on Chromatic can stream an emulator’s output to the device. Writing the ROM to a cartridge is a separate step that enables play without the computer.
The cartridge is both the payoff and the constraint
The included activation code enables Chromatic Developer Mode on the user’s computer. ModRetro requires a firmware update and activation before cartridge writing; OpenAI’s deployment instructions also require a compatible rewritable homebrew cartridge and a USB data cable.
OpenAI warns that installing a new ROM erases the game already on the cartridge and can erase its saves. The installer does not back up the cartridge. This makes the physical cartridge a reusable destination for a finished build, with a practical cost to overwriting it.
The official listing presents the protective carrying case and extra Codex Cartridge as additional accessories. That differs from the secondary launch report’s description of an included case; OpenAI explicitly promises one programmable cartridge. The resale-focused launch report also describes public units as final sale, without refunds.
ModRetro also limits the supported creation workflow to original, user-created ROMs. Its quickstart places responsibility for third-party intellectual-property rights on users, says Developer Mode is used at their own risk, and says it does not review or endorse their content.
OpenAI’s guide shows sequential emulator frames for inspecting a playtest. Credit: OpenAI. Source: OpenAI Developers.
Restock and repeatable results are the next tests
Availability is the immediate question. The earlier launch-day report inferred made-to-order production from units remaining available after launch, but that was an inference, not a confirmed manufacturing policy. OpenAI’s later archived restock notice establishes an expected date; it does not establish a production cap or guarantee supply.
For would-be creators, the decision is whether cartridge play justifies the hardware and setup beyond a browser preview. OpenAI showcases a tile-matching puzzle, a dungeon crawler and a rhythm game made with Codex, but examples do not establish a typical development time or success rate. Its own guide leaves making a game fun to repeated prompting and playtesting. Evidence of required revisions and reliable cartridge installation would make that buying decision easier; the published guides establish a workflow, while leaving the effort required to make a satisfying game unresolved.
Cloud Pathways lets customers coordinate large TPU workloads from one Python client, extending the system Google used for Gemini training. General availability on GKE follows earlier customer access; substantial reported usage and internal recovery results still leave customer cost savings unmeasured.
NVIDIA’s GR00T 2 predicts video and robot actions together to transfer skills across bodies. Company tests show gains from human video, but selected tasks, corrective data and recovery agents qualify its strongest success rates.
Anthropic says its new AI model produced circuit boards, enclosures and firmware for three prototypes that needed no electrical fixes. The demonstration avoided factory revisions, according to the company, while leaving certification, manufacturing scale and total costs unproven.
Meta is reportedly building a consulting channel for corporate AI sales after announcing an enterprise platform led by former MongoDB CEO Chirantan Desai. Its hiring remit also covers advertising and messaging, while prospective partners already work with rival AI suppliers.
PwC US plans a December launch for a critical-thinking and judgment certificate, with completion targeted by June 2027. It will assess staff through performance reviews as the firm seeks new ways to develop skills once practiced through routine work.
Linnea Quigley and Beverly Randolph say Aftermath departed from the musical they agreed to make, exposing a gap between a marketed cast reunion and performers’ control over their screen identities.
Anthropic’s Claude Fable 5.2 and Mythos 5.2 bring lower token prices and stronger technical performance. The release also limits capabilities that could accelerate frontier AI development, as new evidence puts the pace of research close to a sensitive policy boundary.
Meta’s smallest Muse model matches Claude Haiku 5.5’s rounded intelligence score at about one-third of the benchmark task cost. The case for delegating agent work depends on caching, task difficulty and higher charges for long requests.
SpaceX proposes a gas-fired power plant and super-compute facility on about 444 acres in Starbase, Texas. A December hearing will consider local incentives, while the public notice leaves the tax concession and operating demands unspecified.
The California state senator elected to Nancy Pelosi’s open seat wants a federal AI regulator, independent safety checks and penalties for developers. His California record shows both the reach and limits of that agenda; political support from an Anthropic-funded group adds scrutiny.
Apple Creator Studio subscribers can buy extra Logic Pro assistant usage for $49.99 in the U.S. The optional packs last up to a year, but Apple’s public support document does not quantify their capacity.
Boston Dynamics has fitted its electric Atlas humanoid with scooter wheels to test learned controls against real motor limits. The research feeds into the walking robot as Hyundai plans factory deployment, but the demonstration offers no quantified reliability gains.
Nominal raised $240 million at a $3 billion valuation, three times its March valuation, to expand its hardware engineering software into AI agents and lifecycle applications. Its defense foothold and company-reported workflow gains provide a foundation, but do not establish the newer agents’ performance or economics.
The former Megaspeed executive sought backing for a data centre project worth several hundred million dollars, citing GPU access and Alibaba business. The reported pitch adds commercial detail to scrutiny of her ties to Nvidia’s chief, while espionage and chip-smuggling allegations remain unproven.
President Donald Trump offered an unnamed executive’s alleged reversal as evidence that AI risks can be managed. His account identifies neither the conduct at issue nor the legal basis for a jail threat, leaving its value as a safety precedent unproven.
The Texas startup is building factory-made data centers for AI customers, with testing reportedly planned for early 2027. Its financing backs a manufacturing strategy whose cost and power-access advantages remain to be demonstrated.
Baidu says its automated research system has improved inference for the forthcoming ERNIE-5.5-Flash, alongside public benchmark gains. The results show bounded engineering progress, while the cost of finding those improvements remains undisclosed.
Nvidia’s shares fell as other chip stocks rebounded after its board opened an independent inquiry into reports of a foreign intelligence effort targeting CEO and co-founder Jensen Huang. The board retains confidence in him and says it has no reason to believe secrets were compromised; China rejects the spying allegation.
