Runway Gen-3 Retired: Pricing and Workflow Changes

Runway Gen-3 Alpha and Gen-3 Alpha Turbo are retired. Runway's official Gen-3 notice gives July 8, 2026 for Alpha and July 30, 2026 for Turbo. An old tutorial or pricing page is not evidence that you can still select either model in Runway.

This guide was corrected on October 22, 2026. It previously treated surviving documentation as proof of continued availability. That was wrong. The useful material here now concerns interpreting old project records, carrying a shot brief forward, and choosing a replacement by task. It does not tell you to purchase a plan to unlock Gen-3.

Keep the distinction between a model and a platform clear. Runway continues to offer other creative workflows, but a current Runway subscription is not a promise of access to every historical model. The same caution applies to third-party model listings: check the actual provider, version, supported inputs, and working availability before paying for a project.

What replaces each Gen-3 task?

At the time of this correction, Runway's retirement notice points text-to-video and image-to-video users toward Gen-4.5, keyframe work toward Animate Frames, and video-to-video work toward Edit Studio Aleph 2.0. These are task-specific directions from Runway, not a claim that their settings, pricing, or outputs match Gen-3.

Choose the row before choosing a model. A service that makes attractive text-to-video clips may not accept the source footage or endpoint frames required by your project. Confirm those requirements in the current interface and documentation, then run a small trial with material you are authorized to process.

How to read old Gen-3 pricing

Historical subscription prices and credit rates describe a particular billing period and product surface. They cannot establish the current cost of a retired model. This page therefore no longer publishes an active Gen-3 plan recommendation. For a current Runway purchase, review Runway's pricing page and the specific workflow's usage details at the time you decide.

When reconciling an old project, use actual receipts and usage logs. Record the number of attempts, billed duration, any upscaling, and which files were approved. A five-second final clip might have required several rejected attempts. Counting only the final file understates the resources used to produce it.

For the next project, set a cash or allowance limit in advance and stop at that limit. Keep the two measures distinct. A monthly allowance assigned to a clip is an internal planning estimate; a provider invoice is a cash record. Include manual cleanup time when deciding whether a supposedly cheaper generation route really saves effort.

Rebuild the shot brief before rewriting the prompt

Separate the old project's creative decisions from its model controls. Your creative brief can survive a retirement. Its implementation settings may not. Save the approved reference, final framing, intended motion, delivery requirements, and acceptance notes in a folder an editor can understand without reopening the old app.

• Subject: the exact person, object, or location that must stay recognizable.
• Action: one event a viewer can follow inside the clip.
• Camera: the minimum motion needed to communicate that event.
• Protected details: features that cannot change, such as packaging, wardrobe, geometry, or a doorway.
• Delivery: ratio, length, audio needs, caption area, and how the shot joins the next one.

Then translate the brief into the replacement's documented input pattern. Do not paste a legacy control name into prose and assume the model will obey it. If a control exists as a real setting, set it there. If it does not exist, simplify the shot or choose a workflow that exposes the necessary input.

A product-motion example you can adapt

Suppose the job is a short reveal of a plain reusable bottle. This is an illustrative planning example, not a reported benchmark. The bottle should remain stationary, preserve its cap and proportions, and occupy enough of the frame for a mobile viewer to recognize it. The simplest first test uses a slow camera push rather than simultaneous bottle rotation, moving water, and changing lighting.

Do not begin by requesting multiple cuts. First establish that the model can hold the object through one movement. Watch for changes to the neck, disappearing seams, a sliding base, or reflections that contradict the light. A sharp image can still be unsuitable if it changes the thing you sell.

If the bottle drifts, compare a locked camera version against the moving version using the same reference. If the locked version also fails, motion may not be the cause. Check the input crop, ambiguity in the silhouette, and the workflow's reference controls. Save both failures with notes so the next attempt addresses a specific defect.

Validate the transition to production

A successful trial needs more than one attractive frame. Check the complete output at normal speed and frame by frame near difficult moments. Confirm that the intended ratio survives export, there is room for the real caption, and the first and last frames can join the surrounding edit without a distracting jump.

Keep narration and music requirements explicit. Do not assume every replacement generates audio, preserves source audio, or accepts dialogue. If audio must be produced separately, budget that step and test synchronization before committing to a longer sequence. Use the video quality-control checklist for the final review.

Record all attempts, total usage, accepted clips, and repair time. Compare two candidates on the same brief rather than different showcase scenes. An inconclusive test is a reason to narrow the shot or gather better evidence, not to declare a universal winner.

How QuestStudio fits the updated workflow

Video Lab provides a place to inspect currently exposed model choices and prepare a supported shot. Check the selected model's inputs and usage requirements before generating. This page does not promise access to retired Gen-3 models through QuestStudio.

Use Image Lab when you need to prepare a reference image, and Prompt Lab to organize the brief and revisions. Assemble the final sequence in your timeline editor. The goal is a repeatable, reviewed deliverable whose requirements are not tied to one retired model.

Frequently asked questions