Create without counting credits.

Generations on eligible open-source models are not metered or charged individually. Fair use covers a single person or household. Vibe-coded application development and agentic use are allowed and encouraged, but multi-user production applications and workloads with continuous, 24/7 demand are not fair use and may be terminated. For larger Enterprise needs, contact [email protected], or use pay-as-you-go Premium Spark for minimal usage restrictions and maximum throughput. Users who violate these rules or act in bad faith will have their access terminated and will be banned from future use of the Sogni platform. Capacity, queue, and abuse-prevention controls also apply, including unpublished daily fair-use throughput tiers that adjust with GPU supply and network demand. Paid plans include a daily fair-use capacity on the Fast network: reaching it pauses Fast plan rendering until your capacity resets 24 hours later, on your own plan schedule, while Relaxed rendering stays unlimited and Premium Spark is unaffected. Available concurrency also steps down as an account crosses each higher usage threshold. Your current standing is always visible as a percentage at app.sogni.ai/usage. Standard MiniMax H3 sees the most dynamic throttling because of demand, but fewer than 10% of Unlimited subscribers currently experience a fair-use over-usage slowdown. Within each priority class, renders favor accounts with the least usage that UTC day. Premium Spark and SOGNI pay-as-you-go jobs receive priority over subscription jobs, and Unlimited Pro receives priority over Unlimited.

Unlimited covers 200+ supported open-source models, including examples such as MiniMax H3, Z-Image Turbo, Krea 2 Turbo, Qwen Image and Image Edit, Chroma1-HD, Wan 2.2, LTX-2.5, and ACE-Step. Seedance, HappyHorse 1.1, Wan 3, Wan 3 Uncensored, and GPT Image 2 are private vendor models that use separately purchased Premium Spark. Unless a release is specifically marked otherwise, newly supported open-weight models and related updates are automatically included for existing and new subscribers. See the full model coverage table for every model and variant, including those that require Premium Spark.

Choose FastH3 Turbo for most projects. Built on FastVideo FastH3 4-step Preview v1, it renders up to 6× faster than Standard on a full 15-second clip, so a 15-second video can finish in about two minutes once rendering starts, making FastH3 our go-to for fast drafts, quick iterations, and everyday creation; the familiar LightX2V Turbo is one switch away. Choose Standard H3 when you want the absolute best detail and sound; a 15-second Standard video currently averages 13–15 minutes. Both create complete 5- to 15-second videos with dialogue, music, and sound built in. Unlimited includes both under fair use, with one Standard video at a time on Unlimited and two on Unlimited Pro. Video draws on the same daily fair-use capacity as other covered models, and Relaxed rendering has none, so you can keep adding work to your queue throughout the day. Prefer pay as you go? FastH3 costs 4 Spark ($0.02) per output second at both 480p and 544/768p-class output; an exact 8-second, 192-frame clip costs 32 Spark ($0.16). At 480p, Standard costs $0.05 per second, Balanced costs $0.03, and LightX2V Turbo costs $0.02. At 544/768p, Standard costs $0.08, Balanced costs $0.05, and LightX2V Turbo costs $0.03. Ref2VA input video costs $0.05 per second at 480p or $0.08 at 544/768p, even with Turbo output. Sogni plans to add 2K as soon as MiniMax releases the upgrade. Speeds and plan terms are subject to change as Sogni balances supply and demand to keep renders fast for artists and rewarding for the people who share their GPUs through our people-powered render network. Sogni Unlimited is the best deal in town for frequent creators, and we plan to keep it that way. See the MiniMax H3 model guide for workflow and reference-input details.

Creative Agent and API access are included. The generations it dispatches are covered when it uses supported open-source models. If a workflow uses a model marked Premium, that model's separate Premium Spark pricing applies. Individual creators can keep substantial queues moving throughout the day; use the Sogni Creative Agent Skill for Claude, Codex, and Hermes for automated creative workflows. Unattended continuous 24/7 infrastructure and multi-user production workloads are not fair use.

Yes. Unlimited subscribers receive 5% off GPT Image 2, Seedance, HappyHorse 1.1, Wan 3, and Wan 3 Uncensored. Unlimited Pro subscribers receive 10% off those named partner models. They remain pay-as-you-go and are not covered generations.

A payment method is required to begin. Unless you downgrade to Free before the trial ends, billing starts automatically after three days: US$20/month or US$199/year for Unlimited, US$50/month or US$498/year for Unlimited Pro, plus applicable taxes. Only one trial is permitted per household. Annual billing saves 17% and is charged once for the full term. Prefer crypto? You can prepay any plan in USDC or USDT at dashboard.sogni.ai — prepaid plans never auto-renew and have no free trial.

In Sogni Wallet, choose Update Plan, then Downgrade to Free. Review what changes, then continue to your billing provider to confirm. Downgrade a Stripe/web trial before it ends to avoid the first subscription charge; your trial access continues until its original end date. For an active paid plan, access continues until the date shown by the billing provider. Apple and Google Play subscriptions follow the store's cancellation terms. Crypto-paid plans have nothing to cancel — they are prepaid and never renew, so access simply ends on your paid-through date. See the step-by-step downgrade guide and the Cancellation and Refund Policy.

No. Sogni runs in the browser and handles the compute through its community GPU network. API access is also included for supported workflows.