Code2Video-AA

An implementation of HeyGen's Code2Video Bench, evaluating how well models create short product and brand videos from 114 design briefs using HyperFrames HTML.

Models: 22 current, 28 in total
Median: 1025
Range: 304–1536

Score distribution

Lowest: 304
Median: 1025
Highest: 1536

Leaderboard

Include deprecated models

Rank	Model	Score
1	Claude Opus 5.5 (Max, Default Fallback) Anthropic	1536
2	Claude Sonnet 5.5 (Max, Default Fallback) Anthropic	1443
3	GPT-6 Astra (Max) OpenAI	1427
4	GPT-6.1 Sol (Max) OpenAI	1369
5	Gemini 4 Argon (High) Google	1259
6	Claude Haiku 5.5 (Max) Anthropic	1192
7	Grok 4.7 (Xhigh) SpaceXAI	1127
8	MiMo-V2.6-Flash Xiaomi	1052
9	MiMo-V2.6-Pro Xiaomi	1052
10	GLM-5.3 (Max) Z AI	1034
11	GPT-6 Luna (Max) OpenAI	1030
12	Qwen3.8 Max (0902) Alibaba	1021
13	Inkling 2 Small Thinking Machines	1020
14	Muse Wonder (Max) Meta	997
15	GLM-5.3-Flash Z AI	975
16	Qwen3.8 27B (Xhigh) Alibaba	949
17	Gemini 3.8 Flash (High) Google	899
18	Step 5 Preview StepFun	821
19	MiniMax-M3 MiniMax	724
20	Beam 1.0 Reflection	675
21	Inkling (Xhigh) Thinking Machines	467
22	Nemotron 3 Ultra 550B A55B (Reasoning) NVIDIA	304

1
Claude Opus 5.5 (Max, Default Fallback)
Anthropic
1536
2
Claude Sonnet 5.5 (Max, Default Fallback)

Data from Artificial Analysis

Command Palette

Code2Video-AA

Score distribution

Leaderboard