fal.ai

Available

fal.ai has 15 models tracked on Modelglass across 3 verticals (Image, Video, 3D), most recently added 2026-06.

fal.ai has 4 image models (4 current-gen), 7 video models (7 current-gen), and 4 3d models on Modelglass, most recently added 2026-06.

fal.ai ↗

Coverage across verticals

Image 4 models

4 current-gen

Image pricing page ↗

Inference host for many open and partner models. Often prices per output megapixel rather than per image, so the same model can cost differently here than on other hosts.

Video 7 models

7 current-gen

Video pricing page ↗

fal.ai is a model hosting platform that provides serverless inference for many video generation models. Prices are per-second or per-video (per_clip) depending on the model. fal.ai acts as a reseller/host — the model creator and pricing may differ from native API rates.

3D 4 models

5 ungraded

3D pricing page ↗

Inference aggregator. Hosts 3D-generation models (both open-weights and partner-hosted) at flat per-generation USD prices with no subscription. Tracked endpoints: fal-ai/hyper3d/rodin (Rodin); fal-ai/hunyuan3d/v2 and fal-ai/hunyuan3d-v21 (Hunyuan3D v2.x); fal-ai/hunyuan-3d/v3.1/{pro,rapid} (Hunyuan3D 3.1); hitem3d/hi3d/image-to-3d (Hi3D, $0.02/credit like the first-party API); fal-ai/trellis and fal-ai/trellis-2 (Microsoft TRELLIS). Meshy V7 is a private/partner endpoint (no public per-generation price). Pricing is per output; some endpoints add a texture/PBR multiplier or add-on (Hunyuan3D v2 3x for textured mesh; Hunyuan3D 3.1 +$0.15 each for PBR / multi-view / custom face count; Rodin "HighPack" 3x for 4K + high-poly).

See how fal.ai's models fit into a cost-aware routing setup

Release cadence

15 models tracked across all verticals. Earliest release 2024-08, most recent 2026-06.

Derived from Modelglass's per-vertical provider registries — image/llm/video/audio's registry/providers/*.yaml records, plus science/coding/agentic/gdpval's benchmark registries, where a provider is identified by the prefix before "/" in each model's model_id rather than a registry record of its own. Where the same company uses a different provider slug or model_id prefix across verticals (e.g. Google DeepMind vs. Google Cloud, or Qwen's "alibaba" model_id prefix in the coding registry vs. its "qwen" provider slug elsewhere), coverage is merged onto a single profile per an explicit, human-reviewed alias list (registry/provider-aliases.yaml) — everything else reflects exact-slug or exact-prefix matches only.