DeepSeek-V4-Flash leads overall
Ahead on 1 of 2 benchmarks, averaging 13.6 points higher
Summarised from the 2 benchmarks both models were scored on; details in the charts below. A further 1 Elo/rating-scale benchmarks are left out of the average — their scale cannot be added to percentages.
DeepSeek-V4-Flash
DeepSeek-AI
Gemini 3.5 Flash-Lite
Google Deep Mind
Updates live with the mode filters below.
Best overall
DeepSeek-V4-Flash · 563.67
Best single
Gemini 3.5 Flash-Lite · Creative Writing 1556.00
Modality coverage
Gemini 3.5 Flash-Lite · 4 modalities
Head to head
3
Benchmarks
1
Wins
2
Losses
+8.93
Average diff
Benchmark-by-benchmark comparison. Changing the thinking mode or tool filters updates the chart and table below.
Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology
Every model and runtime mode, benchmark by benchmark. Values are comparable along a row, not between different benchmarks.
3 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.
| Benchmark | DeepSeek-V4-Flash | Gemini 3.5 Flash-Lite |
|---|---|---|
SWE-Bench Pro - Public 编程与软件工程 | 52.60Thinking Level · Extra High ｜ Tools | 54.20Thinking Enabled ｜ Tools |
Creative Writing 写作和创作 | 1555.70Standard Mode | 1556.00Standard Mode |
Terminal-Bench 2.1 AI Agent - 工具使用 | 82.70Thinking Level · High ｜ Tools | 54.00Thinking Enabled ｜ Tools |
Official list prices per model API, split by input and output. Unit: USD per 1M tokens.
Architecture, licensing and API modalities. "Not provided" means the field is missing from our database.
| Features & specs | DeepSeek-V4-FlashDeepSeek-AI | Gemini 3.5 Flash-LiteGoogle Deep Mind |
|---|---|---|
Core specsRelease | 2026-04-24 | 2026-07-21 |
Context length | 1M | 1M |
Total parameters | 284B | — |
Active parameters | 13B | N/A |
Max output length | 384,000 tokens | 65,536 tokens |
Architecture | MoE (mixture of experts) | Dense |
Runtime modes | 最高关闭高 | 低中高 |
LicenseCode Open Source | Open Source · MIT License | Closed Source |
Weights Open Source | Open Source · MIT License | Closed Source |
Licensing status | 免费商用授权 | 不开源 |
Local deploymentWeights | Hugging Face | Not provided |
API 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 Flash-Lite |
