Model comparison
DeepSeek V4 Pro vs Qwen3.5 397B
Head-to-head evidence from 12 shared benchmark results across 4 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.
Public leaderboard positions: DeepSeek V4 Pro #46 (Supported); Qwen3.5 397B #71 (Estimated). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.
Evidence parity. DeepSeek V4 Pro and Qwen3.5 397B share 12 comparable benchmark results. 4 of 8 categories are comparable. 11 results are unique to DeepSeek V4 Pro; 43 to Qwen3.5 397B.
Updated July 23, 2026- Shared results
- 12
- DeepSeek V4 Pro only
- 11
- Qwen3.5 397B only
- 43
- Comparable categories
- 4 / 8
Pick DeepSeek V4 Pro if you want the stronger benchmark profile. Qwen3.5 397B only becomes the better choice if mathematics is the priority.
Confidence note. This is a partial-evidence comparison with 12 shared benchmark results across 4 evidence categories; 4 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.
Why this result
DeepSeek V4 Pro is clearly ahead on the BenchAlign aggregate, 60.66 to 57.01. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
DeepSeek V4 Pro's sharpest advantage is in agentic, where it averages 59.1 against 56.5. The single biggest benchmark swing on the page is HMMT Feb 2026, 31.7% to 87.9%. Qwen3.5 397B does hit back in mathematics, so the answer changes if that is the part of the workload you care about most.
Qwen3.5 397B is also the more expensive model on tokens at $0.60 input / $3.60 output per 1M tokens, versus $0.43 input / $0.87 output per 1M tokens for DeepSeek V4 Pro. That is roughly 4.1x on output cost alone. DeepSeek V4 Pro gives you the larger context window at 1M, compared with 128K for Qwen3.5 397B.
Category breakdown
Exact category averages are shown below. Not measured means BenchLM does not have enough sourced public coverage for that model and category.
| Category | DeepSeek V4 Pro | Δ | Qwen3.5 397B |
|---|---|---|---|
| Math | DeepSeek V4 Pro31.7 | Margin→ 58.9 | Qwen3.5 397B90.6 |
| Knowledge | DeepSeek V4 Pro41.3 | Margin→ 15.3 | Qwen3.5 397B56.6 |
| Agentic | DeepSeek V4 Pro59.1 | Margin← 2.6 | Qwen3.5 397B56.5 |
| Coding | DeepSeek V4 Pro65.3 | Margin→ 1.2 | Qwen3.5 397B66.5 |
| Reasoning | DeepSeek V4 ProNot measured | MarginNo overlap | Qwen3.5 397B63.2 |
| Multilingual | DeepSeek V4 ProNot measured | MarginNo overlap | Qwen3.5 397B84.7 |
| Multimodal | DeepSeek V4 ProNot measured | MarginNo overlap | Qwen3.5 397B79.6 |
| Inst. Following | DeepSeek V4 ProNot measured | MarginNo overlap | Qwen3.5 397B92.6 |
Decisive benchmark drivers
The largest measured benchmark gaps in this matchup, with exact reported values.
More
- Source ↗
HMMT Feb 2026
MathA 31.7%B 87.9%Winner: Qwen3.5 397BΔ 56.2HMMT Feb 2026: DeepSeek V4 Pro scored 31.7%; Qwen3.5 397B scored 87.9%. Qwen3.5 397B wins this benchmark. - Source ↗
HLE
KnowledgeA 7.7%B 28.7%Winner: Qwen3.5 397BΔ 21HLE: DeepSeek V4 Pro scored 7.7%; Qwen3.5 397B scored 28.7%. Qwen3.5 397B wins this benchmark. - Source ↗
GPQA
KnowledgeA 72.9%B 88.4%Winner: Qwen3.5 397BΔ 15.5GPQA: DeepSeek V4 Pro scored 72.9%; Qwen3.5 397B scored 88.4%. Qwen3.5 397B wins this benchmark. - Source ↗
Terminal-Bench 2.0
AgenticA 59.1%B 52.5%Winner: DeepSeek V4 ProΔ 6.6Terminal-Bench 2.0: DeepSeek V4 Pro scored 59.1%; Qwen3.5 397B scored 52.5%. DeepSeek V4 Pro wins this benchmark. - Source ↗
MMLU-Pro
KnowledgeA 82.9%B 87.8%Winner: Qwen3.5 397BΔ 4.9MMLU-Pro: DeepSeek V4 Pro scored 82.9%; Qwen3.5 397B scored 87.8%. Qwen3.5 397B wins this benchmark.
Operational comparison
Runtime and commercial metrics are compared only when both models have a complete sourced value.
| Metric | DeepSeek V4 Pro | Qwen3.5 397B | Comparison |
|---|---|---|---|
| Input / output priceUSD per 1M tokens | DeepSeek V4 Pro$0.435 input / $0.87 output | Qwen3.5 397B$0.6 input / $3.6 output | DeepSeek V4 Pro has the lower combined listed price. |
| Generation speedtokens per second | DeepSeek V4 ProNot available | Qwen3.5 397B96 tok/s | A complete speed comparison is not available. |
| First-answer latencyseconds to first token | DeepSeek V4 ProNot available | Qwen3.5 397B2.44 s | A complete latency comparison is not available. |
| Context windowmaximum listed tokens | DeepSeek V4 Pro1M | Qwen3.5 397B128K | DeepSeek V4 Pro lists the larger context window. |
Benchmark Deep Dive
AgenticDeepSeek V4 Pro wins18 benchmarks
| Benchmark | DeepSeek V4 Pro | Qwen3.5 397B | Result |
|---|---|---|---|
| Terminal-Bench 2.0Source | 59.1% | 52.5% | DeepSeek V4 Pro leads |
| MCP AtlasSource | 69.4% | 46.1% | DeepSeek V4 Pro leads |
| ToolathlonSource | 46.3% | 36.3% | DeepSeek V4 Pro leads |
| Claw-EvalSource | 59.8% | 56.8% | DeepSeek V4 Pro leads |
| Gert LabsSource | 50.28% | 46.76% | DeepSeek V4 Pro leads |
| ResearchClawBenchSource | 17.1% | 14.2% | DeepSeek V4 Pro leads |
| BrowseCompSource | — | 62% | Not comparable |
| QwenClawBenchSource | — | 51.8% | Not comparable |
| τ³-bench resultsSource | — | 68.4% | Not comparable |
| VITA-BenchSource | — | 43.7% | Not comparable |
| DeepPlanningSource | — | 37.6% | Not comparable |
| MCP-TasksSource | — | 74.2% | Not comparable |
| WideResearchSource | — | 74.0% | Not comparable |
| τ²-bench resultsSource | — | 95.6% | Not comparable |
| AA Agentic IndexSource | — | 19.9% | Not comparable |
| APEX-Agents-AASource | — | 15.3% | Not comparable |
| GDPval-AASource | — | 23.1% | Not comparable |
| GDPval-AASource | — | 962 | Not comparable |
CodingQwen3.5 397B wins7 benchmarks
| Benchmark | DeepSeek V4 Pro | Qwen3.5 397B | Result |
|---|---|---|---|
| SWE-bench VerifiedSource | 73.6% | 76.2% | Qwen3.5 397B leads |
| SWE-bench ProSource | 52.1% | 50.9% | DeepSeek V4 Pro leads |
| SWE MultilingualSource | 69.8% | — | Not comparable |
| Terminal-Bench 2.0Source | 59.1% | — | Not comparable |
| LiveCodeBench v6Source | — | 83.6% | Not comparable |
| AA-SciCodeSource | — | 42.0% | Not comparable |
| AA Coding IndexSource | — | 48.2% | Not comparable |
Reasoning6 benchmarks
KnowledgeQwen3.5 397B wins15 benchmarks
| Benchmark | DeepSeek V4 Pro | Qwen3.5 397B | Result |
|---|---|---|---|
| MMLU-ProSource | 82.9% | 87.8% | Qwen3.5 397B leads |
| SimpleQASource | 45% | — | Not comparable |
| Chinese-SimpleQASource | 75.8% | — | Not comparable |
| GPQASource | 72.9% | 88.4% | Qwen3.5 397B leads |
| GPQA-DSource | 72.9% | — | Not comparable |
| HLESource | 7.7% | 28.7% | Qwen3.5 397B leads |
| SuperGPQASource | — | 70.4% | Not comparable |
| MMLU-ReduxSource | — | 94.9% | Not comparable |
| C-EvalSource | — | 93% | Not comparable |
| Artificial Analysis Intelligence IndexSource | — | 33.7% | Not comparable |
| AA-GPQA DiamondSource | — | 89.3% | Not comparable |
| AA-HLESource | — | 27.3% | Not comparable |
| AA-Omniscience IndexSource | — | -29.8% | Not comparable |
| AA-Omniscience AccuracySource | — | 31.4% | Not comparable |
| AA-Omniscience Hallucination RateSource | — | 89.1% | Not comparable |
MathQwen3.5 397B wins8 benchmarks
| Benchmark | DeepSeek V4 Pro | Qwen3.5 397B | Result |
|---|---|---|---|
| HMMT Feb 2026Source | 31.7% | 87.9% | Qwen3.5 397B leads |
| IMOAnswerBenchSource | 35.3% | — | Not comparable |
| ApexSource | 0.4% | — | Not comparable |
| Apex ShortlistSource | 9.2% | — | Not comparable |
| AIME26Source | — | 93.3% | Not comparable |
| HMMT Feb 2025Source | — | 94.8% | Not comparable |
| HMMT Nov 2025Source | — | 92.7% | Not comparable |
| MMAnswerBenchSource | — | 80.9% | Not comparable |
Multilingual2 benchmarks
Multimodal8 benchmarks
| Benchmark | DeepSeek V4 Pro | Qwen3.5 397B | Result |
|---|---|---|---|
| Design Arena WebsiteSource | 1264 | — | Not comparable |
| MMMU-ProSource | — | 79% | Not comparable |
| MathVisionSource | — | 88.6% | Not comparable |
| CharXivSource | — | 80.8% | Not comparable |
| VideoMMMUSource | — | 84.7% | Not comparable |
| ScreenSpot ProSource | — | 65.6% | Not comparable |
| V*Source | — | 95.8% | Not comparable |
| AA-MMMU-ProSource | — | 77.3% | Not comparable |
Frequently Asked Questions (5)
Which is better, DeepSeek V4 Pro or Qwen3.5 397B?
DeepSeek V4 Pro is ahead on BenchLM's BenchAlign leaderboard, 60.66 to 57.01. The biggest single separator in this matchup is HMMT Feb 2026, where the scores are 31.7% and 87.9%.
Which is better for knowledge tasks, DeepSeek V4 Pro or Qwen3.5 397B?
Qwen3.5 397B has the edge for knowledge tasks in this comparison, averaging 56.6 versus 41.3. Inside this category, HLE is the benchmark that creates the most daylight between them.
Which is better for coding, DeepSeek V4 Pro or Qwen3.5 397B?
Qwen3.5 397B has the edge for coding in this comparison, averaging 66.5 versus 65.3. Inside this category, SWE-bench Verified is the benchmark that creates the most daylight between them.
Which is better for math, DeepSeek V4 Pro or Qwen3.5 397B?
Qwen3.5 397B has the edge for math in this comparison, averaging 90.6 versus 31.7. Inside this category, HMMT Feb 2026 is the benchmark that creates the most daylight between them.
Which is better for agentic tasks, DeepSeek V4 Pro or Qwen3.5 397B?
DeepSeek V4 Pro has the edge for agentic tasks in this comparison, averaging 59.1 versus 56.5. Inside this category, MCP Atlas is the benchmark that creates the most daylight between them.
Related Comparisons
Explore More
Choose a model with this week’s evidence
Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.
One email each week. Unsubscribe anytime.