Sora Pricing and Active Video Generation Costs

Sora has no price today because OpenAI no longer sells it, through the app or the API.

The closest live price to Sora 2 Standard's last rate is Veo 3.1 Fast, at $0.10 to $0.12 per second.

At Layer3Labs, we build workflow automation for small and mid-sized businesses, and when a model inside a workflow is discontinued, the first job is pricing the same output on a live model.

Sora Pricing Status and Discontinuation Dates

You cannot buy Sora at any price. The web and application interfaces closed on April 26, 2026. The developer Application Programming Interface (API) shut down on September 24, 2026.

OpenAI announced the shutdown on March 24, 2026, as reported by NewsBytes. Because both the application and the programmatic endpoints are offline, no commercial licenses or subscriptions remain for purchase. Readers reviewing the closure schedule can inspect our complete timeline in the guide to whether Sora is shutting down.

Video-model prices change without notice, so verify current figures on the Gemini API pricing page and the BytePlus ModelArk pricing page before budgeting. Teams that previously planned infrastructure budgets around Sora must now transition to active commercial alternatives.

Historical API Rates for Sora 2

Sora 2's last published API rates, as reported, ran from $0.10 to $0.70 per second of video depending on tier and resolution. As reported by third-party tracking platform CostGoat, OpenAI maintained distinct pricing tiers for Standard and Pro access before removing public documentation.

Sora 2 Standard cost $0.10 per second at 720p resolution. Sora 2 Pro cost $0.30 per second at 720p, $0.50 per second at 1024p (1792×1024), and $0.70 per second at 1080p. We have not verified what Sora app credits cost or which ChatGPT plans included Sora, so we quote no figure for either.

No one can pay these Sora 2 rates now. Use them only to compare what the same clip costs on a live model.

• Sora 2 Standard: $0.10 per second at 720p resolution
• Sora 2 Pro: $0.30 per second at 720p resolution
• Sora 2 Pro: $0.50 per second at 1024p resolution (1792×1024)
• Sora 2 Pro: $0.70 per second at 1080p resolution

Credit Balances and Refund Options

According to OpenAI's help page, credits you bought for ChatGPT or Sora can still be used for Codex.

For refunds, OpenAI's help page points to the standard ChatGPT subscription refund instructions.

OpenAI's help page describes an export route: go to sora.chatgpt.com/sunset, click Export, and OpenAI emails you when the export is ready. The same page says OpenAI will permanently delete data associated with Sora after any final export window, so check OpenAI's help page before relying on that route.

Per-Clip Cost of Sora Replacements

Veo 3.1 costs $0.05 to $0.60 per second in the Gemini API, and Seedance 2.5 costs $0.103 to $0.569. The main pay-per-second replacements for Sora are Google Veo 3.1, Gemini Omni Flash, and ByteDance Seedance. Google sells Veo to developers through the Gemini API.

Google provides three preview variants of Veo 3.1. Veo 3.1 Standard costs $0.40 per second at 720p and 1080p, and $0.60 per second at 4K. Veo 3.1 Fast costs $0.10 per second at 720p, $0.12 at 1080p, and $0.30 at 4K. Veo 3.1 Lite costs $0.05 per second at 720p and $0.08 at 1080p. Audio is natively generated and included on all Veo generations. Detailed pricing structures appear in our Veo 3 pricing guide.

Google's pricing documentation also lists Gemini Omni Flash, billed at 5,792 output tokens per second of 720p video. Google states this equals approximately $0.10 per second. BytePlus sells ByteDance's Dreamina Seedance series on ModelArk. BytePlus pricing records show Seedance 2.5 at $1.156 for a 5-second 720p clip ($0.231 per second) and $2.843 for a 5-second 1080p clip ($0.569 per second). Seedance 2.0 Mini lists at $0.08 per second at 720p ($0.38 for 5 seconds). Enterprise accounts receive a discount between August 7 and October 7, 2026, dropping Seedance 2.0 Mini to approximately $0.03 per second.

Each line below multiplies 8 seconds by the model's per-second rate. Spec and workflow differences are in our breakdown of Sora vs Veo and our direct evaluation of Seedance vs Sora vs Veo.

• 8-second 720p on Veo 3.1 Lite: 8 × $0.05 = $0.40
• 8-second 720p on Veo 3.1 Fast: 8 × $0.10 = $0.80
• 8-second 720p on Gemini Omni Flash: 8 × $0.10 = approximately $0.80
• 8-second 720p on Seedance 2.0 Mini list: 8 × $0.08 = $0.64
• 8-second 720p on Seedance 2.5: 8 × $0.231 = approximately $1.85
• 8-second 1080p on Veo 3.1 Fast: 8 × $0.12 = $0.96
• 8-second 1080p on Veo 3.1 Standard: 8 × $0.40 = $3.20
• 8-second 1080p on Seedance 2.5: 8 × $0.569 = approximately $4.55
• 8-second 1080p on Sora 2 Pro, last published rate as reported (no longer sold): 8 × $0.70 = $5.60

Free Generation Tiers and Evaluation Options

Neither Veo 3.1 nor Gemini Omni Flash has a free tier in the Gemini API, according to the Gemini API pricing page.

Teams seeking no-cost evaluation can test hosted creative suites like Runway. Runway's pricing documentation specifies a free tier that grants 125 one-time credits. Paid plans begin at $15 per month billed monthly, or $12 per month billed annually, for 625 monthly credits. Runway allows users to run models such as Gen-4.5 at 60 credits per 5 seconds, Aleph 2.0 at 140 credits per 5 seconds, and Seedance 2.0 Pro 1080p at 160 credits per 4 seconds.

Kling AI, Hailuo and Higgsfield also sell video generation, but we have not verified their prices or free credits, so we give no figure for them. Additional comparisons appear in our analysis of Kling AI pricing and our guide to the best AI video generators.

Selection Criteria for Sora Pricing Alternatives

Choosing an alternative video model depends on clip duration limits, aspect ratio requirements, and developer onboarding terms. Google's video documentation advises developers to treat Gemini Omni Flash as their default generation tool. Google recommends Veo 3.1 for specific capabilities: scene extension, last-frame control, or integration with legacy pipelines.

Clip duration constraints vary substantially across platforms. Veo 3.1 generates clips of 4, 6, or 8 seconds per call, requiring 8 seconds for 1080p or 4K. It supports extension up to 148 seconds at 720p in 7-second increments. Google does not state how extension seconds are billed, so confirm on the Gemini API pricing page before budgeting long clips. BytePlus ModelArk documentation indicates that Seedance 2.5 produces clips from 4 to 30 seconds in a single request. Before you enable Seedance 2.5, BytePlus recommends an account balance above USD 30, a USD 30+ AI Savings Plan, or a Seedance resource pack.

For most work priced on Sora 2 Standard, Veo 3.1 Fast is the pick: the same $0.10 per second at 720p, with audio included. Choose Seedance 2.5 instead if you need single takes longer than 8 seconds and can accept $0.231 per second at 720p. If you need a free way to test, neither fits, so start on Runway's free plan. If OpenAI released a new video API at a comparable per-second rate, we would redo this comparison.