DeepSeek-V4-Pro
DeepSeek- Arena score
- 1451
- Among open models
- #10of 148
- Among all on Arena
- #51
- Votes
- 56,857
Measured on DeepSeek-V4-Pro, the served version, not a local build. LMArena, CC BY 4.0, published Nov 11, 2026.
Score by category
Top 10 among open models in#5Polish#5Creative writing#5Japanese#7French
CategoryScoreOpen rankArena rankVotes
- Chinese1481#21#763.3K
- French1472#7#462.2K
- Coding1469#14#7216.9K
- Software & IT1469#12#6623.3K
- Science1468#7#459K
- Multi-turn1467#7#4210.3K
- Hard prompts, English1466#10#5517.6K
- Medicine & health1465#11#494K
- Legal & government1464#8#494.4K
- Expert1463#17#725.7K
- Hard prompts1461#10#5938K
- English1459#10#4925K
- Polish1459#5#421.1K
- Long prompts1458#10#5725.5K
- German1457#9#48994
- Spanish1456#11#501.7K
- Applied math1455#17#723.2K
- Russian1454#8#506.2K
- Excluding ties1450#10#5143.1K
- Instructions1448#10#553.4K
- Writing & literature1447#9#4913.9K
- Creative writing1446#5#399.6K
- Japanese1445#5#39716
- Business & finance1443#10#5411.3K
- Math1442#20#803K
- Non-English1439#10#5631.8K
- Media & sports1431#9#4912.6K
- Korean1407#12#631.1K
About as strong, smaller
Compare
Qwen3.8-27BQwen143727.8B

gemma-4-31B-itGoogle144331.3B

Hy3Tencent1439299B
GLM-5.3-FlashZ.ai1472321B
Discussion
1 post this week
- DeepSeek V4 Pro 0813explosion-s · Aug 12, 20261K points
- DeepClaude – Claude Code agent loop with DeepSeek V4 Proalattaran · May 3, 2026678 points
- DeepSeek V4 Pro beats GPT-5.5 Pro on precisionyogthos · Jun 8, 2026397 points
- DeepSeek V4 Pro at 75% off until 31 Maynateb2022 · May 6, 202687 points
- DeepSeek V4 Pro 0813 quietly releasedHiPHInch · Aug 12, 202679 points
Built on it
Ways to run it
- GGUF · llama.cpp, Ollama, LM Studioteamblobfish/DeepSeek-V4-Pro-GGUF5 builds
- MLX · Apple silicon: mlx-lm, LM Studiomlx-community/DeepSeek-V4-Pro2 builds
Most downloaded
BenchClub runs
Not runnable yet
Runs open once a DeepSeek-V4-Pro build is registered.
More from DeepSeek
See more
DeepSeek-V3.29.6M685B

DeepSeek-V4-Flash-07312M304B

DeepSeek-V4.1-Flash1.6M763B

DeepSeek-V4-Flash1.2M291B
Other >300B models
See more
GLM-5.3-FlashZ.ai6.4M321B
InklingThinking Machines1.5M952B
NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16NVIDIA1M561B

Kimi-K3Moonshot AI768.9K2.8T