Pika
Pika

Pika v2.2 Text-to-Video

Advanced text-to-video model supporting up to 10-second videos at 1080p. Features Pikaframes for keyframe transitions, Pikaffects for visual effects, and enhanced creative control tools.
rank
17
cost
$2.10
/min
Pika v2.2 Text-to-Video
Total Score
26
26

Scores

Physics24.0
Prompt Adherence47.0
Animation14.3
2D Animation13.0
3D Animation17.0
Anime Animation13.0
Cinematography32.0
Human24.0
Hands19.0
Animals31.0
Objects31.0
Logic + Consistency25.0
Scene Consistency21.0
Text Fidelity6.0
Actor Performance47.0
Total Score26.0

Evaluation Summary

Good for
  • Prompt adherence
  • Keyframe capabilities
  • Actor performances (average)
Bad for
  • Most categories below average
  • Physics
  • Animation
Summary

Pika v2.2 is a pretty weak model across the board, with no category scoring above average. Its main strength is prompt adherence, so it clearly understands direction and follows prompts well. Actor performances are just about acceptable. The notable feature is keyframe capabilities, which few models have. Otherwise it falls apart in most categories.

Examples

Macro shot: a tattoo artist's gloved hands at work, needle buzzing against skin, other hand stretching the canvas taut. Ink, blood, precision. Five fingers per hand, always.
Cinematic slow-motion of a boxer throwing a punch at a heavy bag. Muscles contract in shoulders and arms, sweat flies, face shows exertion.
A woman practices yoga at sunrise on a cliff overlooking the ocean, flowing from warrior pose into a deep lunge. Wind catches her hair.
A futuristic city skyline at sunset, cinematic drone shot
Close up of a wrinkled elderly woman's weathered hands carefully folding an origami crane. Paper creases with each deliberate fold, fingers moving with practiced precision. Soft warm lighting, intimate macro detail.
A cinematic hummingbird hovering over vivid flowers at sunrise, 8s
close up of a man's face as he sits in a world war one trench, he is overwhelmed by emotion before breaking down into tears.
Tattoo needle writes CARPE DIEM across a woman's shoulder, her head turned in profile. The letters form one at a time, skin reddening realistically.
Blocks are removed one by one by two people, the tower wobbles, tilts and finally collapses, it falls. The removed blocks remain visible nearby.

Compare Models

LTX-2 19B
43
43
Lightricks

LTX-2 19B

Cost/min$2.40
Audio
Realtime
PixVerse v5.5
55
55
Cost/min$4.80
Realtime
Gen-4.5
40
40
Runway

Gen-4.5

Cost/min$3.00
Seedance 1.5 Pro
53
53
Cost/min$12.00
Sora 2 Pro
50
50
Cost/min$12.00
All Comparison Pages
RankModelProviderScoreCompare Page
#1
Seedance 2.0 ProTop 5
ByteDance73Pika v2.2 Text-to-Video vs Seedance 2.0 Pro
#2
Kling 3 ProTop 5
Kling62Pika v2.2 Text-to-Video vs Kling 3 Pro
#3
Kling 2.6Top 5
Kling60Pika v2.2 Text-to-Video vs Kling 2.6
#3
Veo 3Top 5
Google60Pika v2.2 Text-to-Video vs Veo 3
#5
Grok Imagine 1.0Top 5
xai59Pika v2.2 Text-to-Video vs Grok Imagine 1.0
#6
Veo 3.1
Google57Pika v2.2 Text-to-Video vs Veo 3.1
#7
PixVerse v5.5
PixVerse55Pika v2.2 Text-to-Video vs PixVerse v5.5
#7
Veo 2
Google55Pika v2.2 Text-to-Video vs Veo 2
#9
Grok 2025
xai54Pika v2.2 Text-to-Video vs Grok 2025
#10
Seedance 1.5 Pro
ByteDance53Pika v2.2 Text-to-Video vs Seedance 1.5 Pro
#11
Veo 3.1 Fast
Google52Pika v2.2 Text-to-Video vs Veo 3.1 Fast
#12
Sora 2 Pro
OpenAI50Pika v2.2 Text-to-Video vs Sora 2 Pro
#13
Veo 3 Fast
Google48Pika v2.2 Text-to-Video vs Veo 3 Fast
#14
Vidu Q2
Vidu45Pika v2.2 Text-to-Video vs Vidu Q2
#15
LTX-2 19B
Lightricks43Pika v2.2 Text-to-Video vs LTX-2 19B
#16
Gen-4.5
Runway40Pika v2.2 Text-to-Video vs Gen-4.5
#18
Infinity Star
FoundationVision25Pika v2.2 Text-to-Video vs Infinity Star