Is DeepSeek V4 Pro Preview SOTA in Coding › Web dev?#73 in Coding › Web devRanked #73 in Coding › Web dev.
How it scores here
DeepSeek V4 Pro Preview in Coding › Web dev
Primary benchmarks
- LMArena WebDev: 1464 (CI 1457–1470) on Nov 20, 2026 (#86 of 165).
On the category's weighted score it is 53.3 points behind #1 Claude Fable 5.2 and 0.7 behind the next model up, Kimi K2 7 Code.
Nearby on the board
| # | Model | Category score |
|---|---|---|
| 70 | Qwen3 6 Max Preview | 48.9 |
| 71 | Mimo V2.5 Pro | 48.2 |
| 72 | Kimi K2 7 Code | 47.4 |
| 73 | DeepSeek V4 Pro Preview | 46.7 |
| 74 | Qwen3 6 Plus | 46.0 |
| 75 | GPT-5.4 | 45.3 |
| 76 | Gemini 3.1 Pro Preview | 44.5 |
Standings
Coding › Web dev
| # | Model | State |
|---|---|---|
| 1 | Claude Fable 5.2 | Co-SOTA |
| 1 | Claude Opus 5.5 | Co-SOTA |
| 2 | Claude Sonnet 5.5 | Near-SOTA |
| 3 | Gemini 4 Carbon | Near-SOTA |
| 4 | GPT-6.2 Astra | Pending |
| 5 | Gemini 4 Neon | Pending |
| 6 | GPT-6 Astra | Was SOTA |
| 7 | GPT-6.2 Sol | #7 in this category |
| 8 | Qwen4 | #8 in this category |
| 9 | GPT-6.1 Sol | #9 in this category |
| 10 | Claude Fable 5.1 | Was SOTA |
| 11 | Muse Wonder Full | #11 in this category |