第一项始终是图钉的来源。总体可信度是各份资料对上方所用开始和结束时间支持程度的加权平均；资料每比最新的一份旧 180 天，权重减半。
Estimated: the press release is datelined Beijing, 24 September 2025 ("today announced"), ScriptByAI[4] gives 23 September and Kling's 2 October release says the model was "released less than a week ago".
9 October 2025 hands-on: a 'very impressive tool', strong text-to-video and image-to-video, but it would not perform a requested somersault and added unrequested breathing sound.[1]
2 October 2025 Kling release: tops both Artificial Analysis video arenas ahead of Veo 3 and Luma Ray 3; 5 and 10 second clips up to 1080p at 25 credits per 5 seconds.[1]
Dates Kling 2.5 Turbo to 23 September 2025 and cuts the 1080p five-second price from 35 to 25 credits.[1]
有遗漏或错误吗？用你自己的话说明：能佐证此图钉的链接、不同的开始或结束日期及理由，或缺失、有误的信息。AI 会对照此图钉的来源进行核实，搜索更好的来源，并添加任何支持你说法的页面。图钉自身的来源仍然最重要。AI 也会查看图片：显示的是别的东西或显示效果差的图片会被移到后面或替换。
| 2.5 Turbo Pro, 5 s |
|---|
| $0.35 (USD) |
| Each additional second | $0.07 (USD) |
|---|
Price as listed on the fal.ai endpoint page, read 24 Oct 2026; same figures on the image-to-video endpoint. · 来源
| Comparison | Win-loss ratio |
|---|---|
| Text-to-video vs Seedance 1.0 mini | 285% |
| Text-to-video vs Veo3-fast | 212% |
| Text-to-video vs Seedance 1.0 | 160% |
| Image-to-video vs Seedance 1.0 mini | 208% |
| Image-to-video vs Veo3-fast | 289% |
| Image-to-video vs Seedance 1.0 | 164% |
Kling's internal test with invited professionals voting on pairs; not independent. Kling also reports 2.5 Turbo ranked No.1 in Artificial Analysis's text-to-video and image-to-video arenas on 2 Oct 2025 (no Elo figure given). · 来源
I will admit that it’s a very impressive tool as of right now.
Kling's 2.5 / 3.0 line has excellent motion coherence — weight, momentum and object interactions tend to hold together where Veo occasionally softens.
Kling AI’s 2.5 Turbo Video model outperformed Google’s Veo 3 and Luma Labs’ Ray 3 to secure the No.1 spot in both the text-to-video and image-to-video leaderboards
