Hedra
  • Developers
  • Studio
  • Enterprise
  • Blog
Log inSign Up
Open Hedra
Your account
    Explore
  • Home
  • Developers
  • Studio
  • Enterprise
  • Blog
    Log inSign Up
    Open Hedra

Sora is shutting down. Move to a full visual agent.

OpenAI closed the Sora app on April 26, 2026 and retires the Sora API on September 24, 2026. Hedra Agent and Workspaces plan, generate and edit across the leading video models, on GPU infrastructure Hedra runs itself, with post-training for teams that need a model of their own.

Get an API keyRead the docs
POST /v3/models/kling-v3
export HEDRA_API_KEY="<key_id>:<secret>"

curl -X POST \
  "https://api.hedra.com/v3/models/kling-v3" \
  -H "Authorization: Key $HEDRA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "input": {
      "prompt": "A slow dolly through a rain-lit Tokyo alley at night",
      "aspect_ratio": "16:9",
      "resolution": "1080p",
      "duration_ms": 5000,
      "quality": "pro"
  } }'
1080p · 5s · 16.8¢/second · $0.84 per clip

A more modern way to work

Skip the model aggregators. Hedra Agent and Workspaces give your team an autonomous agent that supercharges your workflows with a superintelligent understanding of the latest models: brief it in a shared space and it plans, generates, and edits across video, image, speech, and music, with your team and your agents working alongside it.

Hedra Workspaces home: a composer that says start with an idea, research and generation templates, and a row of shared spaces
Built onSPACES · AGENTS · VIDEO · IMAGE · SPEECH · MUSIC · EDIT VIDEO

GPU infrastructure, an inference engine, and post-training research

Hedra operates GPU clusters and the inference engine that runs diffusion and autoregressive media models on them 8x faster. The same research team post-trains models for customers: supervised fine-tuning, LoRA training, and GRPO, on your data.

An abstract microchip architecture animation
Runs onNVIDIA CUDA · AMD ROCm · GOOGLE TPU · AWS TRAINIUM · CUSTOM ASICS

Hedra and Sora: focus at a glance

Sora was one lab's video model and app, now being retired. Hedra is the infrastructure and research company that serves every lab's models and trains yours.

Comparison of Hedra and Sora by focus area
DimensionHedraSora
What it isHedraGPU infrastructure, inference and post-training researchSoraOpenAI's video app and API, being retired
Core technologyHedraHedra's inference engine, its own GPU clusters, and post-training (SFT, LoRA, GRPO)SoraSora 2 and Sora 2 Pro
ModelsHedraOpen-weight models on Hedra's engine, models post-trained for you, and partner models through a gatewaySoraIts own models; app closed April 26, 2026, API closes September 24, 2026
How you connectHedraAPI, SDK, CLI, MCP, and Hedra WorkspacesSoraSora app and API, until shutdown
Where inference runsHedraHedra's GPU clusters, or your ownSoraOpenAI's cloud
Best fitHedraTeams building visual products who need inference, capacity and custom modelsSoraTeams that need a migration path

Where Sora workloads go: video models on Hedra infrastructure

Rates per second of generated video at the resolution shown. Open-weight models run on Hedra's engine; partner models are reached through the gateway.

Models available through Hedra and their rates
ProviderModel IDRate
KlingModel IDkling-v3Rate16.8¢/s · 1080p
GoogleModel IDveo-31Rate39.29¢/s · 1080p
ByteDanceModel IDseedance-25Rate79.09¢/s · 1080p
MiniMaxModel IDminimax-h3Rate6¢/s · 768p
LLumaModel IDluma-ray-32Rate40¢/s · 1080p
AlibabaModel IDwan-3-0Rate20¢/s · 1080p
KlingModel IDkling-26-proRate14¢/s · 1080p

FAQs

What is happening to Sora?
OpenAI announced on March 24, 2026 that it is shutting Sora down. The web and mobile app closed on April 26, 2026, and the Sora API is retired on September 24, 2026. OpenAI has told users to export their videos before the cutoff.
Is Sora 2 Pro still available on Hedra?
Sora 2 Pro is OpenAI's model, reached through Hedra's gateway only while OpenAI's API is live. It leaves the catalog when the API closes on September 24, 2026. Move production to one of the models above before then; the request shape is the same, only the model id changes.
Can Hedra post-train a model for us?
Yes. Hedra runs supervised fine-tuning, LoRA training, and GRPO on models you own or license, on your data and against your evals, then serves the result on the same engine. The weights are yours.
Ocean Waves at Sunset — Kling 2.1 Master

Infrastructure for visual intelligence.

Get an API keyRead the docs
Hedra
Hedra

Product

Developer PlatformStudioEnterpriseSovereignPricing

Resources

Agent documentationDeveloper documentationBlogUse CasesModelsFeedbackChangelogStatus

Company

AboutCareersContactBuilder programSupportAlternatives

Legal

Privacy PolicyTerms of useAcceptable useCookie PolicyBiometric data policy
LinkedinInstagramDiscord
support@hedra.comHedra 2026 — All rights reserved