DROPGoogle launches Gemini 3.8 Flash TTS and Flash-Lite TTS///DROPGemini Omni brings 1080p generation to Google Vids///BREAKClaude Opus 5.5 is here: 1M context, $4 input / $20 output///BREAKOpenAI adds GPT-6 Sol and Luna to the GPT-6 family///PAPEROpus 5.5's system card documents gains—and evaluation limits///BREAKGrok 4.7 launches at $2 input / $6 output per million tokens///DROPGrok Voice Transcribe 2.0 targets difficult real-world audio///TRENDGoogle steers new Gemini API projects to 3.8 Flash or 3.5 Flash-Lite///DROPGoogle launches Gemini 3.8 Flash TTS and Flash-Lite TTS///DROPGemini Omni brings 1080p generation to Google Vids///BREAKClaude Opus 5.5 is here: 1M context, $4 input / $20 output///BREAKOpenAI adds GPT-6 Sol and Luna to the GPT-6 family///PAPEROpus 5.5's system card documents gains—and evaluation limits///BREAKGrok 4.7 launches at $2 input / $6 output per million tokens///DROPGrok Voice Transcribe 2.0 targets difficult real-world audio///TRENDGoogle steers new Gemini API projects to 3.8 Flash or 3.5 Flash-Lite///
DARC.DEV
All guides
ComparisonLevel · Start#Video#Creator#Gemini

AI video in September: track the model, the product and the rollout

A production checklist built around current Gemini Omni, TTS and creator-tool workflows: usable shots, continuity, rights, narration and delivery.

DCDariusz Czechowicz23 Sept 20263 min readSource briefing
AI video in September: track the model, the product and the rollout
Source-checked Sources & corrections
2 practical sections · checklist · sources

Source briefing. Provider claims are attributed to the sources below. This page does not claim a new independent DARC benchmark.

For: Curious readers and builders getting oriented

⚡ TL;DR — the 30-second version

  • Omni in Vids and TTS are separate product rollouts.
  • Do not present coming-soon narration as available now.
  • Measure usable output and retries, not just generated seconds.

Google's 23 September announcement brings Gemini Omni generation into Google Vids, with 1080p output, clip-duration controls and scene extension. The same announcement describes Flash-Lite TTS narration as coming soon. Those two availability statements should not be merged into a single feature list.

The new Gemini 3.8 Flash TTS and Flash-Lite TTS models support different creative and volume-oriented use cases. Voice replication requires consent verification. Audio watermarking and credentials are part of the release's provenance story; they do not replace checking the rights to the source material.

“A video stack is a production workflow, not a leaderboard. Evaluate continuity, editing control, voice rights and delivery quality together.”
— Editorial interpretation

Test a small shot list relevant to your output: an object close-up, a camera move, a character transition and a shot with text. Record retries, source inputs, credit consumption and usable seconds. Include editing and sound work in the production budget.

This article is a current release briefing and an evaluation checklist. It does not claim that DARC ran a blind comparison of five video models or verified the quality of every provider's latest release.

Field guide / 01

Use a shot plan instead of one giant prompt

Define each shot's purpose, camera movement, subject continuity and usable duration. Generate short candidates and judge them in sequence. A striking isolated shot may break continuity when placed next to the previous one.

Record reference assets, model or router configuration and the selected output. Keep rejected candidates in the cost record so the budget reflects what it took to get a publishable sequence.

Field guide / 02

Audio and rights belong in the production checklist

Check narration accuracy, timing, music rights, voice consent and disclosure. A lip-synced clip with an incorrect number or an unlicensed voice is not ready to publish.

Evaluate exported media at the actual delivery resolution and aspect ratio, including captions and mobile safe areas. Keep future product features labeled as forthcoming instead of building a production promise around a launch-page roadmap.

Before you ship

0/5

Use this as a review aid. Checks reset when you leave this page.

Questions that come up

Should I select a single best video model?

Not necessarily. Your workflow may use different tools for generation, editing and speech. Compare the usable result and total effort for your own shot types.

Put it into practiceReview current creator tool profiles

Sources & verification

Last checked 23 Sept 2026. Publication dates and access conditions belong to the linked originals.

  1. 01Gemini Omni in Google VidsGoogle↗
  2. 02Gemini 3.8 text-to-speech says helloGoogle↗
  3. 03Runway — model routers for video, image and audioRunway↗
  4. 04Midjourney — current model versionsMidjourney↗
  5. 05ElevenLabs — Scribe v2 and optional featuresElevenLabs↗
Editorial policy and correction log →
#Video#Creator#Gemini#Audio