Text-to-video (T2V) models are highly sensitive to prompt specificity, while iterative prompt optimization typically improves generation quality only through repeated video generation and vision-language model evaluation at inference time. We propose TIDE, a framework that distills this costly search process into an interpretable and reusable prompt-rewriting Skill. TIDE formulates prompt rewriting as a set of source-conditioned rule modules, each targeting a distinct semantic dimension. Once distilled, the frozen Skill activates the relevant rules and generates both positive and negative prompts in a single LLM call, followed by only one T2V generation, without test-time visual search or backbone fine-tuning. Evolved using only 100 prompts over 14 hours, TIDE achieves prompt-optimization performance comparable to computationally expensive search-based methods. Moreover, when transferred to previously unseen video-generation models, TIDE outperforms existing approaches specifically designed or trained for those models. Code is available at https://github.com/xiningin/TIDE_T2V.
Figures & tables
Figure 1: Pipeline of the proposed TIDE framework. (1) Prompt Rewriting : Given a raw prompt—typically a short user query—the rewriter activates relevant Skill rules to produce an enhanced prompt. (2) Trajectory Generation : The generated video is critiqued, suggestions guide revised prompt candidates, and tournament selection chooses the next-round incumbent; this loop repeats to iteratively improve generation quality. (3) Experience Distillation : Prompt-optimization trajectories yield reusable, dimension-specific patches for the corresponding rule pools. Once full, a pool’s patches update the associated rule and synchronize it with the Skill for subsequent rewriting.
Method
SC
MS
DD
AQ
IQ
OC
MO
HA
SR
Raw
95.64
98.55
87.50
61.25
67.79
28.97
78.91
81.25
81.59
VISTA
96.60
98.70
75.00
61.28
70.67
27.61
98.83
93.75
82.89
Ours init
95.35
98.59
79.17
60.73
65.78
28.22
92.45
97.92
82.94
Ours distil
96.22
98.63
77.08
61.07
66.43
28.24
91.28
100.0
88.43
Table 1: Comparison of TIDE with multi-round iterative prompt optimization across the nine VBench2 dimensions. The average scores of Raw, VISTA, Oursinit , and Oursdistil are 75.72, 78.37, 77.91, and 78.60, respectively. Oursinit denotes the Skill before trajectory distillation, while Oursdistil denotes the final Skill distilled from VISTA’s optimization trajectories. The best and second-best results are highlighted in bold and underlined, respectively.
VBench2
T2VCompBench
Method
Avg.
SC
MS
DD
AQ
IQ
OC
MO
HA
SR
GN
OI
CogVideoX1.5-5B
Raw
54.83
96.18
99.07
16.20
56.66
62.35
24.45
49.95
72.00
49.83
33.95
42.50
Baseline
60.73
94.31
98.94
33.33
60.24
61.65
26.35
58.82
83.67
58.23
34.80
57.75
VPO
61.58
95.11
98.66
34.26
58.31
60.36
26.06
61.94
92.33
50.20
39.12
61.00
Ours
61.87
95.50
98.96
34.26
56.39
59.56
25.62
68.67
91.00
51.75
42.96
55.90
Table 2: Comparison with other prompt optimization methods across CogVideoX1.5 and Wan2.2.
VBench2
T2VCompBench
Method
Avg.
SC
MS
DD
AQ
IQ
OC
MO
HA
SR
GN
OI
Random Selection
66.76
94.94
98.94
31.94
62.12
64.42
27.07
77.57
96.67
73.81
35.34
71.54
w/o Rule Selection
66.81
94.92
98.65
30.09
63.31
65.36
27.60
72.26
97.00
78.54
34.99
72.21
w/o NP Distil
65.46
94.89
99.25
21.76
61.06
61.75
27.18
79.60
93.33
74.59
35.10
71.55
w/o Anti-Starvation
66.97
94.63
98.90
35.65
62.23
64.08
26.96
78.23
96.00
75.65
34.18
70.12
Ours
67.30
96.11
99.03
24.07
62.91
65.40
26.97
81.02
97.33
73.61
40.62
73.14
Table 3: Ablation results for rule selection, negative-prompt distillation, and anti-starvation scheduling on Wan2.2.
Figure 2: Qualitative comparison of prompt optimization. Left: Ours and VPO on CogVideoX1.5. Right: Ours and Custom on Wan2.2.