- 🎬 Video generation model from MiniMax's H3 Max tier
- 🎯 "Recast" variant focused on reworking existing footage
- 🏢 Built by MiniMax, a Chinese multimodal AI lab
- 🎨 Suited to stylistic reinterpretation and creative video output
- 🧩 Part of a broad H3 family spanning text-, image- and reference-to-video
MiniMax is an AI company building generative models across multiple modalities, with a focus that spans both language understanding and audio creation. Their rapid release cadence in early 2026—delivering several new models within just a few months—reflects an ambitious and…
Explore 11 more models by Minimax →MiniMax H3 Max Recast sits in the top "Max" tier of MiniMax's H3 video generation family, arriving in October 2026 as a specialised variant alongside the lab's other H3 Max entry points. Where MiniMax H3 Max covers straightforward prompt-to-video and the image-to-video variant animates a still frame, Recast is the lineage's reworking mode — aimed at taking existing material and recasting it into new form rather than generating from a blank slate.
The H3 family is unusually wide for a video lineup: it spans the baseline H3 models, the H3 Enhanced quality tier, the latency-optimised H3 Max Turbo, reference-conditioned H3 Max R2V, and the camera-control-oriented H3 Max Multi-Angle. MiniMax pairs these with text models in the M series and generative audio work such as MiniMax Music 2.6 and Speech-02 HD.
Best suited to creators who already have source footage or a defined visual concept and want a high-quality transformation of it — restyling, reinterpretation, and iterative creative passes — rather than one-shot generation from a text prompt alone.
This About section is AI-generated from public sources via VeniceStats + Venice inference, with no human editing. It may contain inaccuracies.
Daily usage · MiniMax H3 Max Recast · Organic sellers
| Seller | Health | Rep↓ | Sales | Video | Routing | $ / s | $ / video |
|---|---|---|---|---|---|---|---|
| Venice.ai Proxy 0x1f22…18c9 | failingupstream error · 2.5 h | 83.56 | $2.38 | 12 s | #1 | $0.198 | — |
"Best price" and the seller table are live Antseed catalog data (advertised $/1M tokens — or $ per generated image for unit-billed image models — not settled amounts). Reputation = buyer trust score (0-100, the Antseed SDK's own formula). "Routing" = the SDK's default buyer routing (what the VPR desktop app ships with): a trust ≥ 60 gate on the effective reputation, then cheapest-first among routable sellers; live failover state (per-peer cooldowns) is buyer-side runtime and not included. Model knowledge (TLDR, provider, About) via the VeniceStats enrichment layer. Advertised catalog, not the model used in any specific purchase. "Usage on Antseed" counts only settlements whose buyers share the per-model split on-chain (metadata v2/v3, opt-in), so every usage figure is a lower bound.