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
Each axis is a category average, normalized to a 100-point radar.
Relative edge: 综合评估 +2.7 / Relative gap: 数学推理 -9.0
Relative edge: 数学推理 +9.0 / Relative gap: 综合评估 -2.7
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
DeepSeek-V4-Flash · 72.87
Best single
DeepSeek-V4-Flash · LiveCodeBench 91.60
Modality coverage
Qwen3.6-27B · 3 modalities
Head to head
10
Benchmarks
8
Wins
2
Losses
+3.91
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.
10 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.
| Benchmark | DeepSeek-V4-Flash | Qwen3.6-27B |
|---|---|---|
HLE 综合评估 | 45.10Thinking Level · Extra High | Tools | 24.00Thinking Enabled |
LiveBench 综合评估 | 67.25Standard Mode | 65.56Standard Mode |
MMLU Pro 综合评估 | 86.40Thinking Level · High | 86.20Thinking Enabled |
GPQA Diamond 科学与综合推理 | 88.10Thinking Level · High | 87.80Thinking Enabled |
LiveCodeBench 编程与软件工程 | 91.60Thinking Level · High | 83.90Thinking Enabled |
SWE-bench Multilingual 编程与软件工程 | 73.30Thinking Level · Extra High | Tools | 71.30Thinking Enabled | Tools |
SWE-Bench Pro - Public 编程与软件工程 | 52.60Thinking Level · Extra High | Tools | 53.50Thinking Enabled | Tools |
SWE-bench Verified 编程与软件工程 | 79.00Thinking Level · Extra High | Tools | 77.20Thinking Enabled | Tools |
Terminal Bench 2.0 AI Agent - 工具使用 | 56.90Thinking Level · Extra High | Tools | 59.30Thinking Enabled | Tools |
IMO-AnswerBench 数学推理 | 88.40Thinking Level · High | 80.80Thinking Enabled |
Side-by-side input/output token pricing
Licensing, MoE architecture, and multi-modality support.
| Features & specs | DeepSeek-V4-FlashDeepSeek-AI | Qwen3.6-27B阿里巴巴 |
|---|---|---|
Core specsRelease | 2026-04-24 | 2026-04-22 |
Context length | 1M | 128K |
Parameters | 2840 | 270 |
Active parameters | 130 | 270 |
Max output | 384000 | 16384 |
MoE | Yes | No |
LicenseCode Open Source | Closed Source | Not provided |
Weights Open Source | Closed Source | Not provided |
Commercial use | 免费商用授权 | 免费商用授权 |
Modality supportText Input/Output | / | / |
Image Input/Output | Not provided | / |
Video Input/Output | Not provided | / |
ResourcesPaper / report | DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence | Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model |
DataLearner blog | Not provided | 阿里正式开源Qwen3.6-27B:代码智能体能力上超越全面超越前代旗舰版本之 Qwen3.5-397B-A17B |

Qwen3.6-27B
阿里巴巴