AI news: Grok image API migration, ChatGPT usage plans and Gemini Computer Use — 2 November 2026

Publisher disclosure: MUVO offers competing music, remix, and video tools. This report summarizes public sources and provider documentation; it is not an independent product test or a claim that MUVO is free of similar issues. Read our reporting method.

Key product updates at a glance

Grok’s image-quality API redirect is scheduled for 2 November, and OpenAI now documents shared ChatGPT/Codex usage allowances starting in December. Google’s Computer Use GA and Mistral’s improved Large 4 weights are also included after being held in yesterday’s unpublished draft.

The research window is 30 October at 14:21 through 2 November at 11:59 Europe/Istanbul (UTC+3), with pending findings and scheduled effective dates rechecked. Mistral’s 29 October announcement is carried forward explicitly from earlier research. None of these older announcements is presented as a new launch today. See the separate incident edition for disruptions and recoveries.

Grok image API: the image-quality redirect is due today

The official migration guide sets 2 November as the retirement date for grok-imagine-image-quality. Requests to that slug on image-generation and image-editing endpoints are scheduled to run on grok-imagine-image-2.0 with quality: "low". The notice dates from 2 September. Original migration guide.

The provider says request/response shapes stay compatible and the old slug keeps resolving. Existing grok-imagine-image-pro redirects follow the same route; grok-imagine-image 1.0 is unaffected. The guide states that the replacement’s low-quality rate is $0.01 less per image at every resolution than the retired model.

Practical impact: API developers should review the model and quality settings and the model field returned in their own responses. MUVO has not tested a paid generation or independently verified the switchover at this cutoff. The guide supplies no exact rollout hour or regional schedule; the date is the provider’s effective-date commitment, not a measured deployment time.

OpenAI: shared ChatGPT and Codex usage is planned for December

OpenAI’s help page says Plus, Pro and Business will share a usage allowance across ChatGPT and Codex starting in December. It describes token-based measurement, with usage depending on the selected model and task; Instant chat will not draw from the allowance. Original usage-limit documentation.

The page says combined remaining usage and reset times will appear in Settings → Usage. At a limit, the documented options are waiting for reset, changing plan or buying credits for eligible features. OpenAI expects most users not to wait for resets more often; that is a provider expectation, not a demonstrated outcome.

Status: forthcoming. We verified this page on 2 November; it displayed a relative update label rather than an exact publication timestamp. It specifies December but no exact start day, numeric allowance or complete rollout schedule. The documentation does not establish an equivalent change for every other plan or the separately billed API. Users planning substantial ChatGPT and Codex work should check their account’s final allowance when the transition arrives.

Gemini API: Computer Use is generally available across two APIs

Google’s 30 October release notes make Computer Use generally available in both the Interactions and GenerateContent APIs. This is an API-tool milestone, not an announcement that every consumer Gemini account can control a computer. Official Gemini API release notes.

Google’s guide describes agents that interpret screenshots and return actions for browser, mobile and desktop environments. Developers still implement the environment that receives and executes those actions. The guide documents action-intent information, configurable safety policies and optional screenshot scanning for prompt injection. It also recommends close supervision for important tasks. Computer Use documentation.

Practical impact: teams can build UI-based workflows or application tests using a generally available tool, while retaining responsibility for execution, permissions and review. The release entry does not establish a universal free allowance, a new price or eligibility in every region. Model and account support must be checked for the intended deployment; MUVO has not benchmarked this release.

Mistral Large 4: weights arrive after the API preview

Mistral’s 29 October announcement releases the improved Large 4 weights, also called Le Chonk. That follows the API preview covered in our 7 October edition. Developers can download the weights or use hosted access; the announcement describes further training since the initial preview. Original weight-release announcement, dated Mistral news index.

The current model documentation still labels Large 4 Public Preview. It lists API prices per million tokens of $0.70 input, $0.07 cached input and $2.10 output. Those are token rates, not a fixed per-task price or the cost of operating a downloaded model. Current model documentation and pricing.

Licensing and deployment choices matter alongside availability

The Mistral Ecosystem Open-Weight License is a modified MIT license. It requires a separate commercial license for commercial use by a licensee or affiliate operating a Model-as-a-Service business, subject to its definitions and exceptions. The text distinguishes internal use and embedded end-user product features. Read the actual terms for the intended deployment; downloadable weights do not imply unrestricted model-service resale. Original license.

Mistral describes three reasoning modes—low, medium and high—and reports coding and efficiency improvements. These remain vendor-reported evaluations, not MUVO tests. Analysis: the useful change for developers is greater control over hosting and customization, coupled with a distinct licensing and operational decision. It is not a new music or video generator, and the weight release alone does not establish that the hosted service has left preview.

Sources

• xAI / SpaceXSI: image-quality migration guide and effective date.
• OpenAI: planned ChatGPT and Codex usage limits.
• Google: 30 October Gemini API release notes and Computer Use guide.
• Mistral: weight-release announcement, news index with announcement date, model documentation and license.

We also checked the specialist music, speech, video and remix watchlist. Previously covered Suno Studio, Mureka MP4, Pika H3 Flex, ElevenLabs SFX v3 and Claude Code releases were deduplicated. An updated help page without a verified new product change was not treated as a launch. See the 30 October and 31 October editions for those releases.

MUVO publishes this reporting and offers competing creative tools. See our reporting method.

Update timeline

• 2 November 2026, 11:59 Istanbul / 08:59 UTC: Research cutoff. Added the Grok migration’s effective date and verified OpenAI’s forthcoming December usage policy. Carried Google’s 30 October GA and Mistral’s 29 October weight release from the unpublished 1 November draft, preserving original dates and availability distinctions.