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-Flash
DeepSeek-AI
Gemini 3.5 Flash
Google Deep Mind
Each axis is a category average, normalized to a 100-point radar.
Relative edge: 编程与软件工程 +6.8 / Relative gap: 常识推理 -30.4
Relative edge: 常识推理 +30.4 / Relative gap: 编程与软件工程 -6.8
Method: for each model and benchmark, the chart first averages all scores in the current mode scope instead of taking the best score, then averages those benchmark scores within each category. Only benchmarks with at least two selected models scored are included; missing values are not counted as zero.
Best overall
Gemini 3.5 Flash · 64.72
Best single
Gemini 3.5 Flash · GPQA Diamond 92.80
Modality coverage
Gemini 3.5 Flash · 4 modalities
Head to head
7
Benchmarks
3
Wins
4
Losses
-2.37
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.
7 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.
| Benchmark | DeepSeek-V4-Flash | Gemini 3.5 Flash |
|---|---|---|
HLE 综合评估 | 45.10Thinking Level · Extra High ｜ Tools | 40.20Thinking Level · High ｜ Tools |
LiveBench 综合评估 | 67.25Standard Mode | 75.02Thinking Level · High |
GPQA Diamond 科学与综合推理 | 88.10Thinking Level · High | 92.80Thinking Level · High |
DeepSWE 编程与软件工程 | 54.40Thinking Level · High ｜ Tools | 37.00Thinking Level · Medium ｜ Tools |
SWE-Bench Pro - Public 编程与软件工程 | 52.60Thinking Level · Extra High ｜ Tools | 55.10Thinking Level · High ｜ Tools |
SimpleBench 常识推理 | 46.30Standard Mode | 76.70Standard Mode |
Terminal-Bench 2.1 AI Agent - 工具使用 | 82.70Thinking Level · High ｜ Tools | 76.20Thinking Level · High ｜ Tools |
Side-by-side input/output token pricing
Licensing, MoE architecture, and multi-modality support.
| Features & specs | DeepSeek-V4-FlashDeepSeek-AI | Gemini 3.5 FlashGoogle Deep Mind |
|---|---|---|
Core specsRelease | 2026-04-24 | 2026-06-20 |
Context length | 1M | 1M |
Parameters | 284B | — |
Active parameters | 13B | Not provided |
Max output | 384000 | 65536 |
MoE | Yes | No |
LicenseCode Open Source | Open Source · MIT License | Closed Source |
Weights Open Source | Open Source · MIT License | Closed Source |
Commercial use | 免费商用授权 | 不开源 |
Local deploymentWeights | Hugging Face | Not provided |
Modality supportText Input/Output | / | / |
Image Input/Output | Not provided | / |
Audio Input/Output | Not provided | / |
Video Input/Output | Not provided | / |
ResourcesPaper / report | DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence | Gemini 3.5: frontier intelligence with action |
