Benchmark Almanac

Image-to-WebDev Arena

Claude Opus 5.5 1,752 leads since September 2026
Not in model cards

Users upload a screenshot, wireframe or mockup, two anonymous models build the website in Code Arena, and users vote for the better result. Votes become ratings with bootstrapped confidence intervals.

On October 19th, 2026 the board listed 65 models and 191,568 votes. It is a separate board from the text-prompt Code Arena WebDev board.

Released
April 2026
Kind
Web and app front ends
Level
Everyday
Scored
Ranked by votes
Answers
Open-ended
Tasks
Kept private
Strongest model tested
GPT-6 Astra #3 on the capability index
Last model added
October 2026
Scores refreshed
November 12th, 2026
55
Models scored on its own leaderboard

Claude Opus 5.5 leads with 1,752

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. Every run is in the next chart.

55 models. Scroll sideways for the rest. Bars start at 1,200.

  1. 1,752 Claude Opus 5.5
  2. 1,744 Claude Sonnet 5.5
  3. 1,724 Claude Fable 5.1
  4. 1,722 GPT-6 Astra
  5. 1,696 GPT-6.1 Sol
  6. 1,664 Claude Opus 5
  7. 1,655 GPT-6 Sol
  8. 1,638 Qwen3.8 Max (0902)
  9. 1,637 Muse Spark 1.3
  10. 1,621 Claude Fable 5
  11. 1,616 Qwen 3.8 Max
  12. 1,610 DeepSeek V4.1 Flash
  13. 1,607 Grok 4.6
  14. 1,606 GPT-5.6 Sol
  15. 1,600 Gemini 3.7 Flash
  16. 1,597 MiMo-V2.6 Pro
  17. 1,594 Gemini 3.8 Flash
  18. 1,593 MiMo-V2.6 Flash
  19. 1,589 Grok 4.7
  20. 1,586 Kimi K3
  21. 1,582 GLM-5.3-Flash
  22. 1,579 Claude Opus 4.7
  23. 1,574 Qwen 3.8 27B
  24. 1,566 GPT-6 Luna
  25. 1,562 Grok 4.5
  26. 1,551 Step 5 Preview
  27. 1,545 Gemini 3.6 Flash
  28. 1,544 Claude Opus 4.6
  29. 1,539 GPT-5.6 Terra
  30. 1,539 Muse Spark 1.1
  31. 1,539 Claude Sonnet 5
  32. 1,531 GPT-5.5
  33. 1,530 Claude Sonnet 4.6
  34. 1,522 Kimi K2.6
  35. 1,517 GPT-5.6 Luna
  36. 1,512 Kimi K2.7 Code
  37. 1,506 Gemini 3.5 Flash
  38. 1,505 Seed2.1 Pro
  39. 1,480 Gemini 3.1 Pro
  40. 1,473 MiniMax-M3
  41. 1,473 Qwen 3.6 Plus
  42. 1,451 Gemini 3 Flash
  43. 1,449 Gemini 3.5 Flash-Lite
  44. 1,444 GPT-5.3 Codex
  45. 1,441 GPT-5.4
  46. 1,441 Gemini 3 Pro
  47. 1,439 Kimi K2.5
  48. 1,427 GLM-5V-Turbo
  49. 1,419 GPT-5.1
  50. 1,410 Inkling-Small
  51. 1,392 Nemotron 3.5 Super VL
  52. 1,376 Grok 4.3
  53. 1,362 Mistral Large 3
  54. 1,338 Gemini 3.1 Flash-Lite
  55. 1,285 Gemini 2.5 Pro

Bars start at 1,200.

  1. 1 Claude Opus 5.5 max 1,752
  2. 2 Claude Sonnet 5.5 xhigh 1,744
  3. 3 Claude Fable 5.1 max 1,724
  4. 4 GPT-6 Astra max 1,722
  5. 5 GPT-6.1 Sol max 1,696
  6. 6 Claude Opus 5 max 1,664
  7. 7 GPT-6 Sol max 1,655
  8. 8 Qwen3.8 Max (0902) 1,638
  9. 9 Muse Spark 1.3 max 1,637
  10. 10 Claude Fable 5 high 1,621
  11. 11 Qwen 3.8 Max 1,616
  12. 12 DeepSeek V4.1 Flash max 1,610
  13. 13 Grok 4.6 high 1,607
  14. 14 GPT-5.6 Sol 1,606
  15. 15 Gemini 3.7 Flash high 1,600
  16. 16 MiMo-V2.6 Pro 1,597
  17. 17 Gemini 3.8 Flash high 1,594
  18. 18 MiMo-V2.6 Flash 1,593
  19. 19 Grok 4.7 xhigh 1,589
  20. 20 Kimi K3 max 1,586
  21. 21 GLM-5.3-Flash 1,582
  22. 22 Claude Opus 4.7 high 1,579
  23. 23 Qwen 3.8 27B 1,574
  24. 24 GPT-6 Luna max 1,566
  25. 25 Grok 4.5 1,562
  26. 26 Step 5 Preview high 1,551
  27. 27 Gemini 3.6 Flash high 1,545
  28. 28 Claude Opus 4.6 high 1,544
  29. 29 GPT-5.6 Terra 1,539
  30. 30 Muse Spark 1.1 1,539
  31. 31 Claude Sonnet 5 high 1,539
  32. 32 GPT-5.5 1,531
  33. 33 Claude Sonnet 4.6 1,530
  34. 34 Kimi K2.6 1,522
  35. 35 GPT-5.6 Luna 1,517
  36. 36 Kimi K2.7 Code 1,512
  37. 37 Gemini 3.5 Flash medium 1,506
  38. 38 Seed2.1 Pro 1,505
  39. 39 Gemini 3.1 Pro 1,480
  40. 40 MiniMax-M3 1,473
  41. 41 Qwen 3.6 Plus 1,473
  42. 42 Gemini 3 Flash 1,451
  43. 43 Gemini 3.5 Flash-Lite 1,449
  44. 44 GPT-5.3 Codex 1,444
  45. 45 GPT-5.4 1,441
  46. 46 Gemini 3 Pro 1,441
  47. 47 Kimi K2.5 thinking 1,439
  48. 48 GLM-5V-Turbo 1,427
  49. 49 GPT-5.1 high 1,419
  50. 50 Inkling-Small 1,410
  51. 51 Nemotron 3.5 Super VL nvfp4 1,392
  52. 52 Grok 4.3 1,376
  53. 53 Mistral Large 3 1,362
  54. 54 Gemini 3.1 Flash-Lite 1,338
  55. 55 Gemini 2.5 Pro Jun 2025 1,285

Source: Official Image-to-WebDev Arena leaderboard

#ModelScoreSettingDateSource
1 Claude Opus 5.5 1,752 max Sep 2026 Official leaderboard
2 Claude Sonnet 5.5 1,744 xhigh Sep 2026 Official leaderboard
3 Claude Fable 5.1 1,724 max Sep 2026 Official leaderboard
4 GPT-6 Astra 1,722 max Sep 2026 Official leaderboard
5 GPT-6.1 Sol 1,696 max Sep 2026 Official leaderboard
6 Claude Opus 5 1,664 max Jul 2026 Official leaderboard
7 GPT-6 Sol 1,655 max Sep 2026 Official leaderboard
8 Qwen3.8 Max (0902) 1,638 Sep 2026 Official leaderboard
9 Muse Spark 1.3 1,637 max Sep 2026 Official leaderboard
10 Claude Fable 5 1,621 high Jun 2026 Official leaderboard
11 Qwen 3.8 Max 1,616 Aug 2026 Official leaderboard
12 DeepSeek V4.1 Flash 1,610 max Sep 2026 Official leaderboard
13 Grok 4.6 1,607 high Aug 2026 Official leaderboard
14 GPT-5.6 Sol 1,606 xHigh, Codex Jul 2026 Official leaderboard
15 Gemini 3.7 Flash 1,600 high Aug 2026 Official leaderboard
16 MiMo-V2.6 Pro 1,597 Sep 2026 Official leaderboard
17 Gemini 3.8 Flash 1,594 high Sep 2026 Official leaderboard
18 MiMo-V2.6 Flash 1,593 Sep 2026 Official leaderboard
19 Grok 4.7 1,589 xhigh Sep 2026 Official leaderboard
20 Kimi K3 1,586 max Jul 2026 Official leaderboard
21 GLM-5.3-Flash 1,582 Aug 2026 Official leaderboard
22 Claude Opus 4.7 1,579 high Apr 2026 Official leaderboard
23 Qwen 3.8 27B 1,574 Aug 2026 Official leaderboard
24 GPT-6 Luna 1,566 max Sep 2026 Official leaderboard
25 Grok 4.5 1,562 Jul 2026 Official leaderboard
26 Step 5 Preview 1,551 high Sep 2026 Official leaderboard
27 Gemini 3.6 Flash 1,545 high Jul 2026 Official leaderboard
28 Claude Opus 4.6 1,544 high Feb 2026 Official leaderboard
29 GPT-5.6 Terra 1,539 xHigh, Codex Jul 2026 Official leaderboard
30 Muse Spark 1.1 1,539 Jul 2026 Official leaderboard
31 Claude Sonnet 5 1,539 high Jun 2026 Official leaderboard
32 GPT-5.5 1,531 xHigh, Codex Apr 2026 Official leaderboard
33 Claude Sonnet 4.6 1,530 Feb 2026 Official leaderboard
34 Kimi K2.6 1,522 Apr 2026 Official leaderboard
35 GPT-5.6 Luna 1,517 xHigh, Codex Jul 2026 Official leaderboard
36 Kimi K2.7 Code 1,512 Jun 2026 Official leaderboard
37 Gemini 3.5 Flash 1,506 medium May 2026 Official leaderboard
38 Seed2.1 Pro 1,505 Jun 2026 Official leaderboard
39 Gemini 3.1 Pro 1,480 Feb 2026 Official leaderboard
40 MiniMax-M3 1,473 Jun 2026 Official leaderboard
41 Qwen 3.6 Plus 1,473 Mar 2026 Official leaderboard
42 Gemini 3 Flash 1,451 Dec 2025 Official leaderboard
43 Gemini 3.5 Flash-Lite 1,449 Jul 2026 Official leaderboard
44 GPT-5.3 Codex 1,444 Codex Feb 2026 Official leaderboard
45 GPT-5.4 1,441 Mar 2026 Official leaderboard
46 Gemini 3 Pro 1,441 Nov 2025 Official leaderboard
47 Kimi K2.5 1,439 thinking Jan 2026 Official leaderboard
48 GLM-5V-Turbo 1,427 Apr 2026 Official leaderboard
49 GPT-5.1 1,419 high Nov 2025 Official leaderboard
50 Inkling-Small 1,410 Jul 2026 Official leaderboard
51 Nemotron 3.5 Super VL 1,392 nvfp4 Oct 2026 Official leaderboard
52 Grok 4.3 1,376 Apr 2026 Official leaderboard
53 Mistral Large 3 1,362 Dec 2025 Official leaderboard
54 Gemini 3.1 Flash-Lite 1,338 Mar 2026 Official leaderboard
55 Gemini 2.5 Pro Jun 2025 1,285 Jun 2025 Official leaderboard

More effort adds a median 39 elo

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
Muse Spark 1.3
xhigh max
1,637
Claude Opus 4.7
high
1,579
Claude Opus 4.6
high
1,544
Gemini 3.5 Flash
medium
1,506
Gemini 3 Flash
minimal
1,451
Kimi K2.5
instant
1,439
GPT-5.1
high
1,419

Source: Official Image-to-WebDev Arena leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Claude Opus 5.5 max 1,752
Claude Sonnet 5.5 xhigh 1,744
Claude Fable 5.1 max 1,724
GPT-6 Astra max 1,722
GPT-6.1 Sol max 1,696
Claude Opus 5 max 1,664
GPT-6 Sol max 1,655
Qwen3.8 Max (0902) 1,638
Muse Spark 1.3 max 1,637
Claude Fable 5 high 1,621
Qwen 3.8 Max 1,616
DeepSeek V4.1 Flash max 1,610
DeepSeek V4.1 Flash max 1,609
Grok 4.6 high 1,607
GPT-5.6 Sol xHigh, Codex 1,606
Gemini 3.7 Flash high 1,600
Muse Spark 1.3 xhigh 1,598
MiMo-V2.6 Pro 1,597
Gemini 3.8 Flash high 1,594
MiMo-V2.6 Flash 1,593
Grok 4.7 xhigh 1,589
Kimi K3 max 1,586
GLM-5.3-Flash 1,582
Claude Opus 4.7 high 1,579
Qwen 3.8 27B 1,574
GPT-6 Luna max 1,566
Claude Opus 4.7 1,562
Grok 4.5 1,562
Step 5 Preview high 1,551
Gemini 3.6 Flash high 1,545
Claude Opus 4.6 high 1,544
GPT-5.6 Terra xHigh, Codex 1,539
Muse Spark 1.1 1,539
Claude Sonnet 5 high 1,539
Claude Opus 4.6 1,538
GPT-5.5 xHigh, Codex 1,531
Claude Sonnet 4.6 1,530
Kimi K2.6 1,522
GPT-5.5 High, Codex 1,519
GPT-5.6 Luna xHigh, Codex 1,517
Kimi K2.7 Code 1,512
Gemini 3.5 Flash medium 1,506
Seed2.1 Pro 1,505
GPT-5.5 Codex 1,504
Gemini 3.5 Flash high 1,500
Gemini 3.1 Pro 1,480
MiniMax-M3 1,473
Qwen 3.6 Plus 1,473
Gemini 3 Flash 1,451
Gemini 3.5 Flash-Lite 1,449
GPT-5.3 Codex Codex 1,444
GPT-5.4 1,441
Gemini 3 Pro 1,441
Kimi K2.5 thinking 1,439
Gemini 3 Flash minimal 1,431
GLM-5V-Turbo 1,427
GPT-5.1 high 1,419
Kimi K2.5 instant 1,411
Inkling-Small 1,410
Nemotron 3.5 Super VL nvfp4 1,392
Grok 4.3 1,376
Mistral Large 3 1,362
GPT-5.1 1,349
Gemini 3.1 Flash-Lite 1,338
Gemini 2.5 Pro Jun 2025 1,285

6 models have held number one on the official leaderboard since its release in April 2026

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

Anthropic
Own leaderboard 6 leaders
Claude Opus 5.5 1,752

Source: Official Image-to-WebDev Arena 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
1,200 1,400 1,600 1,800 2025 2026 1,200 1,400 1,600 1,800 2025 2026

Source: Official Image-to-WebDev Arena 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