Benchmark Almanac

DesignArena Video to Website

Muse Spark 1.3 1,250 leads since September 2026
Not in model cards

Models watch a reference video and recreate its content as a working single-file HTML website. Users compare the rendered pages and vote.

One of 40 DesignArena boards. DesignArena ranks entries with Bradley-Terry ratings from user votes.

Kind
Web and app front ends
Level
Everyday
Scored
Ranked by votes
Answers
Open-ended
Tasks
All public
Strongest model tested
Gemini 3.8 Flash #23 on the capability index
Last model added
October 2026
Scores refreshed
November 12th, 2026
19
Models scored on its own leaderboard

Muse Spark 1.3 leads with 1,250

Best score per model from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best.

19 models. Scroll sideways for the rest. Bars start at 1,000.

  1. 1,250 Muse Spark 1.3
  2. 1,241 Kimi K3
  3. 1,240 Muse Spark 1.2
  4. 1,218 Muse Spark 1.1
  5. 1,213 Gemini 3.7 Flash
  6. 1,211 Gemini 3.8 Flash
  7. 1,200 Gemini 3.6 Flash
  8. 1,195 GLM-5.3-Flash
  9. 1,184 Kimi K2.6
  10. 1,181 Gemini 3.5 Flash
  11. 1,179 Kimi K2.5
  12. 1,178 MiniMax-M3.1-Flash
  13. 1,175 Kimi K2.7 Code
  14. 1,170 Qwen 3.6 Plus
  15. 1,163 MiniMax-M3
  16. 1,146 Gemini 3.1 Pro
  17. 1,145 MiMo-V2.5
  18. 1,141 GLM-5V-Turbo
  19. 1,005 Nemotron 3.5 Super VL

Bars start at 1,000.

  1. 1 Muse Spark 1.3 xhigh 1,250
  2. 2 Kimi K3 1,241
  3. 3 Muse Spark 1.2 1,240
  4. 4 Muse Spark 1.1 1,218
  5. 5 Gemini 3.7 Flash 1,213
  6. 6 Gemini 3.8 Flash 1,211
  7. 7 Gemini 3.6 Flash 1,200
  8. 8 GLM-5.3-Flash 1,195
  9. 9 Kimi K2.6 1,184
  10. 10 Gemini 3.5 Flash 1,181
  11. 11 Kimi K2.5 thinking 1,179
  12. 12 MiniMax-M3.1-Flash 1,178
  13. 13 Kimi K2.7 Code 1,175
  14. 14 Qwen 3.6 Plus 1,170
  15. 15 MiniMax-M3 1,163
  16. 16 Gemini 3.1 Pro 1,146
  17. 17 MiMo-V2.5 1,145
  18. 18 GLM-5V-Turbo 1,141
  19. 19 Nemotron 3.5 Super VL 1,005

Source: Official DesignArena Video to Website leaderboard

#ModelScoreSettingDateSource
1 Muse Spark 1.3 1,250 xhigh Sep 2026 Official leaderboard
2 Kimi K3 1,241 Jul 2026 Official leaderboard
3 Muse Spark 1.2 1,240 Aug 2026 Official leaderboard
4 Muse Spark 1.1 1,218 Jul 2026 Official leaderboard
5 Gemini 3.7 Flash 1,213 Aug 2026 Official leaderboard
6 Gemini 3.8 Flash 1,211 Sep 2026 Official leaderboard
7 Gemini 3.6 Flash 1,200 Jul 2026 Official leaderboard
8 GLM-5.3-Flash 1,195 Aug 2026 Official leaderboard
9 Kimi K2.6 1,184 Apr 2026 Official leaderboard
10 Gemini 3.5 Flash 1,181 May 2026 Official leaderboard
11 Kimi K2.5 1,179 thinking Jan 2026 Official leaderboard
12 MiniMax-M3.1-Flash 1,178 Oct 2026 Official leaderboard
13 Kimi K2.7 Code 1,175 Jun 2026 Official leaderboard
14 Qwen 3.6 Plus 1,170 Mar 2026 Official leaderboard
15 MiniMax-M3 1,163 Jun 2026 Official leaderboard
16 Gemini 3.1 Pro 1,146 Feb 2026 Official leaderboard
17 MiMo-V2.5 1,145 Apr 2026 Official leaderboard
18 GLM-5V-Turbo 1,141 Apr 2026 Official leaderboard
19 Nemotron 3.5 Super VL 1,005 Oct 2026 Official leaderboard

5 models have held number one on the official leaderboard since 2023

Lab claims count from the day the card came out, independent runs from the model’s release.

Moonshot AI Meta
Own leaderboard 5 leaders
Muse Spark 1.3 1,250

Source: Official DesignArena Video to Website leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

On the benchmark’s own leaderboard
1,000 1,050 1,100 1,150 1,200 1,250 2026 1,000 1,050 1,100 1,150 1,200 1,250 2026

Source: Official DesignArena Video to Website leaderboard

No lab model card reports it yet

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.

More in Coding

All 282 in Coding
  • SWE-bench Verified In use 500 human-validated GitHub issues from 12 Python repos. Agent writes a patch that must pass hidden tests. Claude Opus 4.7 83.5%
  • LiveCodeBench v6 In use Competitive programming problems collected after model cutoffs from LeetCode, AtCoder, Codeforces. Scored pass@1 on hidden tests. DeepSeek-V4-Pro 93.5%
  • SWE-bench Multilingual In use 300 GitHub-issue tasks across nine programming languages in SWE-bench format. Patch graded by repository tests. Claude Fable 5.2 96.6%
  • DeepSWE In use 113 original long-horizon engineering tasks across 91 repos and five languages, written from scratch. Hand-written verifiers accept alternative solutions. GPT-6 Astra 74.1%
  • SciCode In use Scientist-curated research coding problems split into subproblems across natural sciences. Code checked against test cases. Seed2.1 Pro 59.8%
  • Codeforces rating In use Model solves recent Codeforces contest problems. An Elo-style rating is computed from simulated contest performance. DeepSeek V4.1 Flash 3,471