Best coding models for design

This page uses the WebDev Arena benchmark as a frontend/design signal, then adds current token price, context, and host availability. It is not a general visual-design quality ranking: the benchmark source is an external dataset and remains non-commercial until its upstream licence is confirmed.

#ModelWebDev scoreInput $/MTokOutput $/MTokHostsContext
1Claude Opus 51687.0$5.00$25.0011.0M
2Kimi K31674.3$3.00$15.002—
3Claude Fable 51626.3$10.00$50.0011.0M
4GLM 5.31614.5$2.10$6.6011.0M
5GLM 5.3 Flash1607.1$0.23$0.7511.0M
6Qwen3.8 27b1593.3$0.40$2.373—
7GLM-5.21593.3$1.20$4.204—
8Claude Opus 4.71566.8$5.00$25.0011.0M
9Claude Opus 4.61556.3$5.00$25.0011.0M
10Claude Opus 4.81552.2$5.00$25.0011.0M
11Gemini 3.6 Flash1527.8$0.75$3.7511.0M
12Claude Sonnet 4.61521.0$3.00$15.0011.0M
13Kimi-K2.61508.8$0.73$3.503262k
14GLM 5.11508.3$1.40$4.401—
15Gemini 3.5 Flash1506.4$1.50$9.0011.0M
16MiniMax M31486.7$0.30$1.203—
17Kimi K2.7 Code1472.7$0.75$3.504262k
18DeepSeek V4 Pro1463.6$1.32$3.505—
19GPT-5.51457.8$5.00$30.0011.1M
20Gemini 3.5 Flash Lite1446.6$0.30$2.5011.0M
21Kimi K2.51436.4$0.50$2.807—
22GLM 51435.9$1.20$3.846—
23GLM 4.71434.7$0.72$2.646—
24MiniMax M2.71397.5$0.30$1.201197k
25Minimax M2.11387.5$0.36$1.446—