Vision2Web
Agents build websites from 918 prototype images across 193 tasks, from static pages to full-stack apps. A GUI agent and a VLM judge score them.
- Made by
- Z.ai (zai-org)
- Released
- March 2026
- Kind
- Web and app front ends
- Level
- Expert
- Scored
- Graded by a model
- Answers
- Open-ended
- Tasks
- All public
- Strongest model tested
- Claude Opus 5 #10 on the capability index
- Last in a launch post
- November 2026
- Last model added
- November 2026
- Scores refreshed
- November 12th, 2026
In the authors’ words
Recent advances in large language models have improved the capabilities of coding agents, yet systematic evaluation of complex, end-to-end website development remains limited. To address this gap, we introduce Vision2Web, a hierarchical benchmark for visual website development, spanning from static UI-to-code generation, interactive multi-page frontend reproduction, to long-horizon full-stack website development. The benchmark is constructed from real-world websites and comprises a total of 193 tasks across 16 categories, with 918 prototype images and 1,255 test cases. To support flexible, thorough and reliable evaluation, we propose workflow-based agent verification paradigm based on two complementary components: a GUI agent verifier and a VLM-based judge. We evaluate multiple visual language models instantiated under different coding-agent frameworks, revealing substantial performance gaps at all task levels, with state-of-the-art models still struggling on full-stack development.
Claude Fable 5 leads with 82.3% among models released by August 2026
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.
Claude Fable 5 leads with 82.3%, claims included
Best score per model from every source, including labs’ own numbers and claims Epoch AI collected from reports and papers. Labs pick their own settings, so a claim can beat an independent run of the same model.
The newest model with a score here came out in August 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- Claude Fable 5.2
-
Gemini 4 Carbon - GPT-6 Astra
- Claude Opus 5.5
- and 12 more
- 1 Claude Fable 5 82.3%
- 2 Claude Opus 5 80.2%
- 3 Kimi K3 79%
- 4 GLM-5.3-Flash 77.8%
- 5 Gemini 3.7 Flash 76.5%
- 6 Claude Opus 4.8 76.1%
- 7 GPT-5.6 Sol 73.5%
- 8
DeepSeek-V4-Flash-Vision-Exp 67.6% - 9 GPT-5.6 Terra 67.5%
- 10 GPT-5.6 Luna 64.5%
Source: Official Vision2Web leaderboard
- 82.3 Claude Fable 5
- 80.2 Claude Opus 5
- 79.0 Kimi K3
- 78.6
LongCat-2.5 - 77.8 GLM-5.3-Flash
- 76.5 Gemini 3.7 Flash
- 76.1 Claude Opus 4.8
- 73.5 GPT-5.6 Sol
- 69.0
Qwen 3.8 Max - 67.6 DeepSeek-V4-Flash-Vision-Exp
- 67.5 GPT-5.6 Terra
- 64.5 GPT-5.6 Luna
- 64.0 Qwen3.8 Flash
- 62.9 Qwen 3.8 27B
- 62.9 Qwen3.8-Omni-Flash
- 31.0 GLM-5V-Turbo
- 1 Claude Fable 5 82.3%
- 2 Claude Opus 5 80.2%
- 3 Kimi K3 79%
- 4 LongCat-2.5 78.6%
- 5 GLM-5.3-Flash 77.8%
- 6 Gemini 3.7 Flash 76.5%
- 7 Claude Opus 4.8 76.1%
- 8 GPT-5.6 Sol 73.5%
- 9 Qwen 3.8 Max 69%
- 10 DeepSeek-V4-Flash-Vision-Exp 67.6%
- 11 GPT-5.6 Terra 67.5%
- 12 GPT-5.6 Luna 64.5%
- 13 Qwen3.8 Flash 64%
- 14 Qwen 3.8 27B 62.9%
- 15 Qwen3.8-Omni-Flash 62.9%
- 16 GLM-5V-Turbo 31%
Sources: Official Vision2Web leaderboard and labs’ model cards
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Claude Fable 5 | 82.3% | Claude Code | Jun 2026 | Official leaderboard |
| 2 | Claude Opus 5 | 80.2% | Claude Code | Jul 2026 | Official leaderboard |
| 3 | Kimi K3 | 79% | Claude Code | Jul 2026 | Official leaderboard |
| 4 | GLM-5.3-Flash | 77.8% | Claude Code | Aug 2026 | Official leaderboard |
| 5 | Gemini 3.7 Flash | 76.5% | Gemini CLI | Aug 2026 | Official leaderboard |
| 6 | Claude Opus 4.8 | 76.1% | Claude Code | May 2026 | Official leaderboard |
| 7 | GPT-5.6 Sol | 73.5% | Codex | Jul 2026 | Official leaderboard |
| 8 | DeepSeek-V4-Flash-Vision-Exp | 67.6% | Claude Code | Aug 2026 | Official leaderboard |
| 9 | GPT-5.6 Terra | 67.5% | Codex | Jul 2026 | Official leaderboard |
| 10 | GPT-5.6 Luna | 64.5% | Codex | Jul 2026 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Claude Fable 5 | 82.3% | Claude Code | Jun 2026 | Official leaderboard |
| 2 | Claude Opus 5 | 80.2% | Claude Code | Jul 2026 | Official leaderboard |
| 3 | Kimi K3 | 79% | Claude Code | Jul 2026 | Official leaderboard |
| 4 | LongCat-2.5 | 78.6% | Nov 2026 | Lab model card | |
| 5 | GLM-5.3-Flash | 77.8% | Claude Code | Aug 2026 | Official leaderboard |
| 6 | Gemini 3.7 Flash | 76.5% | Gemini CLI | Aug 2026 | Official leaderboard |
| 7 | Claude Opus 4.8 | 76.1% | Claude Code | May 2026 | Official leaderboard |
| 8 | GPT-5.6 Sol | 73.5% | Codex | Jul 2026 | Official leaderboard |
| 9 | Qwen 3.8 Max | 69% | Aug 2026 | Lab model card | |
| 10 | DeepSeek-V4-Flash-Vision-Exp | 67.6% | Claude Code | Aug 2026 | Official leaderboard |
| 11 | GPT-5.6 Terra | 67.5% | Codex | Jul 2026 | Official leaderboard |
| 12 | GPT-5.6 Luna | 64.5% | Codex | Jul 2026 | Official leaderboard |
| 13 | Qwen3.8 Flash | 64% | Aug 2026 | Lab model card | |
| 14 | Qwen 3.8 27B | 62.9% | Aug 2026 | Lab model card | |
| 15 | Qwen3.8-Omni-Flash | 62.9% | Sep 2026 | Lab model card | |
| 16 | GLM-5V-Turbo | 31% | Apr 2026 | Lab model card |
Labs report scores for 7 models in their cards. Switch the scores above to rank them together.
2 models have held number one on the official leaderboard since its release in March 2026 3 models have held number one with claims included since its release in March 2026
Lab claims count from the day the card came out, independent runs from the model’s release. We draw one row each for the benchmark’s own leaderboard and what labs claim in their cards.
Sources: Official Vision2Web leaderboard and labs’ model cards
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Sources: Official Vision2Web leaderboard and labs’ model cards
3 labs put it in a launch post in the last six months
Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.
| Date | Launch post | Score | As the post writes it |
|---|---|---|---|
| Nov 2, 2026 | LongCat-2.5 | 78.6 | Vision2Web |
| Sep 18, 2026 | Qwen3.8-Omni-Flash | 62.9 | Vision2Web |
| Aug 26, 2026 | Qwen3.8 Flash | 64 | Visual web development Vision2Web |
| Aug 26, 2026 | GLM-5.3-Flash | 77.8 | Vision2Web |
| Aug 14, 2026 | Qwen 3.8 27B | 62.9 | Visual web development Vision2Web |
| Aug 3, 2026 | Qwen 3.8 Max | 69 | Vision2Web |
| Apr 1, 2026 | GLM-5V-Turbo | 31 | Vision2Web 31.0 |
One system card reports it
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| Sep 22, 2026 | Qwen3.8-Omni-Flash technical report | 62.9 | Vision2Web, mean of frontend, webpage and website, Claude Code, GPT-5.4 judge (Table 3) | Capabilities |
One lab claim next to the leaderboard
Each score a lab reported for its own model, beside the same model's best score on the leaderboard. Gaps mostly come from different harnesses, effort levels and task subsets. 1 of them match, which usually means one side copied the other. Every benchmark.
| Model | Lab says | Board says | Gap |
|---|---|---|---|
| GLM-5.3-Flash | 77.8% Vision2Web | 77.8% Leaderboard, Claude Code | Same |
Versions
More 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