| Released | 2 Feb 2026 |
|---|
| Max clip length | 10 seconds |
|---|
| Max resolution | 720p |
|---|
| Audio | xAI: 'dramatically better audio' |
|---|
| Usage figure | xAI: 1.245 billion videos generated in the last 30 days |
|---|
Taken from xAI's 2 Feb 2026 post on X as quoted and embedded by TechSpot; xAI has no news page for 1.0. API figures are on the API launch pin. · 出典
Estimate from the version's published specs and reviews, 0-5; scores are only comparable within one criteria year. Scored as launched (1.0, 2 Feb 2026). Resolution 3: 720p. Duration 4: the apps cap at 10 s per xAI's post but the same model allowed 15 s through the API (DeepLearning.AI), so the longest single clip is 15 s. Audio 5: expressive character voices and synced music, though TechSpot says lip-sync is still imperfect. Control 3: text, one image and input video for editing, no multi-reference input. Speed 2.5: Artificial Analysis 110.1 s per video (length unspecified) against xAI's later '40+ s for a 6 s 720p clip'; midpoint used. Realism 3.5: top of the Artificial Analysis blind-vote board for both modes at launch, but TechSpot and Analytics Vidhya describe visible flaws and lip-sync trouble, and the model was only fourth for text-to-video on Arena; the reviews are mixed press plus one how-to blog, with no tier-1 hands-on comparison.[1]
However, the videos reveal that lip syncing remains a challenge for generative AI.
Grok Imagine came in first in the text-to-video generation category, outperforming high-performance models like Google's Veo 3.1 and OpenAI's Sora 2 Pro.
both the videos clearly show that AI-generated videos are far from being perfect right now, and I believe there is still time before we give them a prompt and stay assured of an error-free, quality output.
Grok Imagine 1.0 debuted at the top of the Artificial Analysis Video Arena, a blind, head-to-head test of preferences judged by human viewers. It’s slower than some competitors but generally less expensive.
読者の写真で、出典付きのピンのメディアとは別に表示されます。食べたものや見たものを評価でき、レストランのピンでは写真上に料理をタグ付けできます。
最初の項目は常にピンの出典です。全体の確度は、各資料が上記の開始・終了時刻をどれだけ強く裏付けているかの加重平均です。最新の資料より180日古くなるごとに、重みは半分になります。
xAI の投稿 (「Grok Imagine 1.0 の紹介、これまでで最大の飛躍」、Gigazine と TechSpot[2] によって埋め込まれています) の日付は 2026 年 2 月 2 日です。 Wikipedia[5] の Grok の記事では 2 月 1 日となっていますが、一部のブログでは 2 月 3 日となっているため、xAI 自身の発表の日付が使用されています。
Confirms 10 s and 720p, shows creators chaining shots into 50-second films, and notes lip-sync remains a challenge and one creator saw clips degrade by 10 s.[1]
Hands-on with two 10-second prompts: appealing for a stand-alone message, good audio, still far from error-free; also says clips were 6 seconds before.[1]
Independent summary of the launch benchmarks: first on Artificial Analysis for text- and image-to-video, first for image-to-video on LM Arena, fourth for text-to-video.[1]
States that on 1 February 2026 xAI released Imagine 1.0 with improved audio quality.[1]
足りない点や誤りがありますか? このピンを裏付けるリンク、別の開始日・終了日とその理由、欠けている情報や誤っている情報を、自由に書いてください。AIがこのピンの出典と照らし合わせ、より良い情報源を探し、裏付けとなるページがあれば参考資料として追加します。ピン自身の出典が引き続き最も重視されます。AIは画像も確認します。別のものが写っている画像や写りの悪い画像は、後ろに回すか差し替えます。
元のタイトル: Grok Imagine 1.0 Released With 10-Second 720p Video