fal launches H3 Max Lip Sync: talking videos from a photo in seconds

fal has officially launched H3 Max Lip Sync: upload a photo plus an audio clip and get a lip-synced talking video in seconds. According to fal's own benchmarks, the model ranks first in both quality and speed, with a median generation time of just 11 seconds; the model is a fine-tuned customization of MiniMax H3 Max.

Confirmed

• Features: one photo plus one audio file produces a lip-synced talking video, with natural expressions and motion, supporting any language.
• Performance: fal's own evaluation ranks it first in quality and speed, with a median generation time of 11 seconds; a post by the development team member isidentical also cites "under 3 seconds," so both figures coexist in official messaging.
• Origin: Arabic-language blogger aziz4ai reports the model is a fine-tuned MiniMax H3 Max on the fal platform, calling it the strongest lip sync solution available, and noting it is very cheap.
• Use cases: fal says the speed supports large-scale rapid iteration across multiple characters, campaigns, and languages.

Unconfirmed

• The "first in quality and speed" claim comes from fal's own benchmarks with no independent third-party verification; the "3 seconds" vs "11 seconds" figures differ, likely depending on scenario or task size.

Why it matters

• Lip sync is a key step in advertising, multilingual localization, and virtual character production; H3 Max's second-level speed and low cost could reshape the cost structure of producing such content at scale.

2026-09-18 ~ 2026-09-19 · 5 related posts

• [source] fal launches H3 Max Lip Sync: photo + audio to talking video in under 3 seconds — isidentical · 2026-09-18
• fal launches H3 Max Lip Sync: photo + audio to lip-synced video in 11 seconds — jfischoff · 2026-09-19
• [source] fal's finetuned MiniMax H3 Max delivers cheap, top-tier lip sync — aziz4ai · 2026-09-19

2 near-duplicate retellings: jfischoff · isidentical