Is DeepSeek V4 Flash Preview SOTA in Coding?Not enough data#106 on LMArena Text (style-controlled, coding); missing Terminal-Bench 4.0 and SWE-bench Pro V2 HARD
How it scores here
DeepSeek V4 Flash Preview in Coding
Primary benchmarks
- LMArena Text (style-controlled, coding): 1480 (CI 1474–1486) on Nov 19, 2026 (#131 of 260).
Standings
Coding
| # | Model | State |
|---|---|---|
| 1 | Claude Sonnet 5.5 | Co-SOTA |
| 1 | Claude Opus 5.5 | Co-SOTA |
| 1 | Kimi K3.1 Pro | Co-SOTA |
| 2 | Claude Fable 5.2 | Near-SOTA |
| 3 | GPT-6 Astra | Near-SOTA |
| 4 | GPT-6.1 Sol | #4 in this category |
| 5 | Claude Fable 5.1 | #5 in this category |
| 6 | Claude Opus 5 | #6 in this category |
| 7 | Gemini 4 Argon | #7 in this category |
| 8 | GPT-6.1 Terra | #8 in this category |
| 9 | GPT-6 Sol | #9 in this category |
| 10 | Muse Wonder | #10 in this category |