See key specs and per-benchmark scores for each model/mode. Scroll horizontally for all columns. 当前对比 2 个模型的评测数据与核心参数。
DeepSeek-V4-Pro
DeepSeek-AI
Best overall
DeepSeek-V4-Pro · 79.98
Best single
DeepSeek-V4-Pro · LiveCodeBench 93.50
Modality coverage
DeepSeek-V4-Pro · 2 modalities
Head to head
5
Benchmarks
5
Wins
0
Losses
+17.06
Average diff
Compare benchmark results across thinking modes and tool usage.
Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology
Complete scores for each model/mode across selected benchmarks.
5 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.
| Benchmark | DeepSeek-V4-Pro | DeepSeek-R1-0528 |
|---|---|---|
GPQA Diamond 综合评估 | 90.10Thinking Level · High | 81.00Thinking Enabled |
HLE 综合评估 | 48.20Thinking Level · Extra High ｜ Tools | 17.70Thinking Enabled |
MMLU Pro 综合评估 | 87.50Thinking Level · High | 85.00Thinking Enabled |
LiveCodeBench 编程与软件工程 | 93.50Thinking Level · High | 73.30Thinking Enabled |
SWE-bench Verified 编程与软件工程 | 80.60Thinking Level · Extra High ｜ Tools | 57.60Thinking Enabled |
Side-by-side input/output token pricing
Licensing, MoE architecture, and multi-modality support.
| Features & specs | DeepSeek-V4-ProDeepSeek-AI | DeepSeek-R1-0528DeepSeek-AI |
|---|---|---|
Core specsRelease | 2026-04-24 | 2025-05-28 |
Context length | 1M | 64K |
Parameters | 16000 | 6710 |
Active parameters | 490 | 370 |
Max output | 384000 | 64000 |
MoE | Yes | Yes |
Supported modes | No mode data | 思考模式（Thinking Mode） |
LicenseCode Open Source | Closed Source | Closed Source |
Weights Open Source | Closed Source | Closed Source |
Commercial use | 免费商用授权 | 免费商用授权 |
Modality supportText Input/Output | / | / |
Image Input/Output | / | Not provided |
ResourcesPaper / report | DeepSeek-V4 Technical Report | DeepSeek-R1-0528 Release |
DeepSeek-R1-0528
DeepSeek-AI
