| Released | 28 Jan 2026 |
|---|
| Offering | Unified bundle of APIs: video generation (text prompt or image) and video editing |
|---|
| Audio | Native video-audio generation |
|---|
| Video editing | Restyle scenes, add/remove/swap objects, control motion and scene |
|---|
| Partner platforms | Also available on partner platforms (ComfyUI and HeyGen quoted) |
|---|
| Latency (P50, 720p, 8 s clip) | Measured end to end by xAI against rivals; the per-model seconds are shown only as a chart, not in text |
|---|
xAI's launch post does not publish max resolution, clip length or per-second price in text; those appear in the second table from the current docs. · sumber
| Spec | Value |
|---|---|
| Grok Imagine vs Kling o1, overall | 57% vs 43% |
| Grok Imagine vs Kling o1, instruction following | 53.1% vs 46.9% |
| Grok Imagine vs Kling o1, consistency | 60.6% vs 39.4% |
| Grok Imagine vs Runway Aleph, overall | 64.1% vs 35.9% |
| Grok Imagine vs Runway Aleph, instruction following | 57.4% vs 42.6% |
| Grok Imagine vs Runway Aleph, consistency | 63.1% vs 36.9% |
Run by xAI's human raters on IVEBench at 1280 x 720; xAI's own comparison, not independent. The post also ranks Grok Imagine first on Artificial Analysis text-to-video (as of 28 Jan 2026), ahead of Veo 3.1 Fast, Veo 3, Sora 2 Pro and Sora 2. · sumber
| Spec | Value |
|---|---|
| API model name | grok-imagine-video |
| Resolution | 480p, 720p (up to 720p) |
| Clip duration | 1-15 s |
| Aspect ratios | 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 21:9, 5:2; image-to-video matches the input image |
| Video editing | Source clip up to 8.7 s |
| Video extension | Adds 2-10 s |
| Audio | Generated audio |
| Price per second, 480p | $0.05 (USD) |
| Price per second, 720p | $0.07 (USD) |
| Input pricing | Image $0.002; video input $0.01 per second |
| Rate limit | 10 requests per second |
Current docs, not necessarily the figures on launch day; the model page says video-extension pricing is a promotional rate subject to change. · sumber
Estimate from the version's published specs and reviews, 0-5; scores are only comparable within one criteria year. Scored as launched (API, 28 Jan 2026): resolution 3 (720p maximum, 480p also offered; DeepLearning.AI). Duration 4: up to 15 s via the API (Rundown, DeepLearning.AI); the apps capped at 10 s five days later. Audio 5: native dialogue, effects and music in the same pass. Control 3: text, a single image, and an input video for prompt-based editing (add/remove/swap objects, restyle); there are no multi-image or audio references yet, and no first/last-frame keyframes, so it sits between the 2 and 4 anchors. Speed 2.5: Artificial Analysis averaged 110.1 s per video (clip length unspecified) while xAI's June 2026 page says its previous model needed '40+ seconds' for a 6 s 720p clip; the midpoint is used. Realism 4 (praised): first on Artificial Analysis for both text- and image-to-video and first for image-to-video on Arena, but fourth for text-to-video on Arena; the reviews are benchmark-led (The Rundown, DeepLearning.AI) plus a reseller (getimg.ai, labelled), with no tier-1 hands-on test.[1]
xAI just released the Grok Imagine API, a new AI video generation and editing suite that jumped to the top of Artificial Analysis rankings for both text and image-to-video outputs while undercutting rivals on price.
Third-party benchmarks show Grok Imagine 1.0 matches or exceeds leading models on quality at a lower cost than premium competitors, a combination that lowers the cost of experimentation.
Grok Imagine’s strongest feature is instruction following in video form
| Grok Imagine by xAI: How It Stacks Up in AI Video Generation · 29 Jan 2026 | 0.70 |
|---|---|
| xAI's Grok Imagine climbs the leaderboards · 30 Jan 2026 | 0.60 |
| xAI’s Cost-Effective Video Generator: Grok Imagine 1.0 sharply cuts costs for high-quality video generation · 27 Mac 2026 | 0.50 |
Foto daripada pembaca, berasingan daripada media bersumber pin. Nilaikan apa yang anda makan atau lihat, dan pada pin restoran, tag hidangan terus pada foto anda.
Entri pertama sentiasa sumber pin. Keyakinan keseluruhan ialah purata berpemberat bagi sejauh mana setiap rujukan menyokong masa mula dan tamat yang digunakan di atas; sesebuah rujukan diambil kira separuh kurang bagi setiap 180 hari lebih lama daripada yang terbaharu.
Halaman berita xAI sendiri bertarikh '28 Jan 2026' dan nota kaki penanda arasnya berbunyi 'sehingga 2026-01-28'; Akhbar teknologi (Rundown AI, 30 Jan) dan halaman model Vercel juga memberikan 28 Januari.
Reports the API handles text-to-video, image-to-video and editing with clips up to 15 seconds and native audio at $4.20 per minute (against $12 for Veo 3.1 and $30 for Sora 2 Pro).[1]
Gives the numbers: first on Artificial Analysis in both text-to-video and image-to-video, 1,280x720 or 854x480 output, up to 15 seconds via the API, average generation time of 110.1 s.[1]
Reseller write-up: strongest on instruction-following and audio with natural lip-sync; label as reseller material.[1]
Current docs give the API duration range as 1-15 seconds, state that the original grok-imagine-video supports up to 720p, and that edited video keeps the original's length (capped at 8.7 s).[1]
Ada yang tertinggal atau salah? Nyatakan dengan kata-kata anda sendiri: pautan yang menyokong pin ini, tarikh mula atau tamat yang berbeza beserta sebabnya, atau fakta yang tiada atau tersilap. AI menyemaknya berbanding sumber pin ini, mencari yang lebih baik, dan menambah mana-mana halaman yang menyokong anda. Sumber pin itu sendiri tetap paling diambil kira. Gambar yang menunjukkan sesuatu yang lain, atau menunjukkannya dengan buruk, turut dilihat, dan dialihkan ke bawah atau diganti.
Tajuk asal: Grok Imagine API Launches With Video Generation and Editing