DeepSeek-V4-FlashvsQwen3.6-27B
Across 10 shared benchmarks, Qwen3.6-27B leads overall: DeepSeek-V4-Flash wins 1, Qwen3.6-27B wins 9, with 0 ties and an average score difference of -12.13.
DeepSeek-V4-Flash
DeepSeek-AI · 2026-04-24 · Reasoning model
Qwen3.6-27B
阿里巴巴 · 2026-04-22 · Reasoning model
DeepSeek-V4-Flash1 win(10%)(90%)9 winsQwen3.6-27B
Benchmark scores
Grouped by capability, sorted by largest gap within each. 10 shared benchmarks.
Coding and Software Engineer
Qwen3.6-27B 4/4| Benchmark | DeepSeek-V4-Flash | Qwen3.6-27B | Diff |
|---|---|---|---|
| LiveCodeBench | 55.2086 / 126Normal (No Tools) | 83.9021 / 126Thinking (No Tools) | -28.70 |
| SWE-Bench Pro - Public | 49.1047 / 57Normal (With Tools) | 53.5036 / 57Thinking (With Tools) | -4.40 |
| SWE-bench Verified | 73.7045 / 114Normal (With Tools) | 77.2028 / 114Thinking (With Tools) | -3.50 |
| SWE-bench Multilingual | 69.7020 / 25Normal (With Tools) | 71.3017 / 25Thinking (With Tools) | -1.60 |
General Knowledge
Qwen3.6-27B 2/3| Benchmark | DeepSeek-V4-Flash | Qwen3.6-27B | Diff |
|---|---|---|---|
| HLE | 8.10164 / 181Normal (No Tools) | 24115 / 181Thinking (No Tools) | -15.90 |
| MMLU Pro | 8346 / 133Normal (No Tools) | 86.2018 / 133Thinking (No Tools) | -3.20 |
| LiveBench | 67.2549 / 115Normal (No Tools) | 65.5652 / 115Normal (No Tools) | +1.69 |
AI Agent - Tool Usage
Qwen3.6-27B 1/1| Benchmark | DeepSeek-V4-Flash | Qwen3.6-27B | Diff |
|---|---|---|---|
| Terminal Bench 2.0 | 49.1036 / 48Normal (With Tools) | 59.3020 / 48Thinking (With Tools) | -10.20 |
General Evaluation
Qwen3.6-27B 1/1| Benchmark | DeepSeek-V4-Flash | Qwen3.6-27B | Diff |
|---|---|---|---|
| GPQA Diamond | 71.20151 / 226Normal (No Tools) | 87.8064 / 226Thinking (No Tools) | -16.60 |
Math and Reasoning
Qwen3.6-27B 1/1| Benchmark | DeepSeek-V4-Flash | Qwen3.6-27B | Diff |
|---|---|---|---|
| IMO-AnswerBench | 41.9022 / 23Normal (No Tools) | 80.8020 / 23Thinking (No Tools) | -38.90 |
Specs
| Field | DeepSeek-V4-Flash | Qwen3.6-27B |
|---|---|---|
| Publisher | DeepSeek-AI | 阿里巴巴 |
| Release date | 2026-04-24 | 2026-04-22 |
| Model type | Reasoning model | Reasoning model |
| Architecture | MoE | Dense |
| Parameters | 284B | 27B |
| Context length | 1M | 128K |
| Max output | 384K | 16K |
API pricing
Prices use DataLearner records when available; missing fields are not inferred.
| Item | DeepSeek-V4-Flash | Qwen3.6-27B |
|---|---|---|
| Text input | $0.14 / 1M tokens | Not public |
| Text output | $0.28 / 1M tokens | Not public |
| Cache read | $0.0028 / 1M tokens | Not public |
One or both models have incomplete public pricing.
Summary
- Qwen3.6-27Bleads in:Coding and Software Engineer (4/4), General Knowledge (2/3), AI Agent - Tool Usage (1/1), General Evaluation (1/1), Math and Reasoning (1/1)
On average across the 10 shared benchmarks, Qwen3.6-27B scores 12.13 higher.
Largest single-benchmark gap: IMO-AnswerBench — DeepSeek-V4-Flash 41.90 vs Qwen3.6-27B 80.80 (-38.90).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.