Model Comparison

DeepSeek-V4-Pro-Max vs Qwen3.5-397B-A17BWhich is better in 2026?

DeepSeek-V4-Pro-Max significantly outperforms across most benchmarks. Qwen3.5-397B-A17B is 1.5x cheaper per token.

Verdict: DeepSeek-V4-Pro-Max vs Qwen3.5-397B-A17B — which is better?

DeepSeek-V4-Pro-Max (by DeepSeek) and Qwen3.5-397B-A17B (by Alibaba Cloud / Qwen Team) are two of the AI models people compare most. Here is how they stack up on benchmarks, price and capabilities, and which one to pick in 2026.

DeepSeek-V4-Pro-Max outperforms in 9 benchmarks (BrowseComp, GDPval-AA, GPQA, Humanity's Last Exam, IMO-AnswerBench, SWE-bench Multilingual, SWE-Bench Verified, Terminal-Bench 2.0, Toolathlon), while Qwen3.5-397B-A17B is better at 1 benchmark (MMLU-Pro). DeepSeek-V4-Pro-Max significantly outperforms across most benchmarks.

On price, Qwen3.5-397B-A17B is roughly 1.5x cheaper per token on a blended 3:1 input/output basis, which adds up quickly at production volume.

DeepSeek-V4-Pro-Max also accepts a larger context window (1,048,576 input tokens), making it the stronger choice for long documents and large codebases.

Choose DeepSeek-V4-Pro-Max if…

  • you want the strongest raw capability — it leads on 9 of 10 shared benchmarks
  • you process long inputs — it offers a 1,048,576 token context window
  • you want the most recent training data — it shipped Apr 2026

Choose Qwen3.5-397B-A17B if…

  • cost matters — it's about 1.5x cheaper per token

Performance Benchmarks

Comparative analysis across standard metrics

10 benchmarks

DeepSeek-V4-Pro-Max outperforms in 9 benchmarks (BrowseComp, GDPval-AA, GPQA, Humanity's Last Exam, IMO-AnswerBench, SWE-bench Multilingual, SWE-Bench Verified, Terminal-Bench 2.0, Toolathlon), while Qwen3.5-397B-A17B is better at 1 benchmark (MMLU-Pro).

DeepSeek-V4-Pro-Max significantly outperforms across most benchmarks.

Fri Jul 24 2026 • llm-stats.com

Arena Performance

Human preference votes

Pricing Analysis

Price comparison per million tokens

Qwen3.5-397B-A17B costs less

For input processing, DeepSeek-V4-Pro-Max ($1.60/1M tokens) is 2.7x more expensive than Qwen3.5-397B-A17B ($0.60/1M tokens).

For output processing, DeepSeek-V4-Pro-Max ($3.20/1M tokens) is 1.1x cheaper than Qwen3.5-397B-A17B ($3.60/1M tokens).

In conclusion, DeepSeek-V4-Pro-Max is more expensive than Qwen3.5-397B-A17B.*

* Using a 3:1 ratio of input to output tokens

Lowest available price from all providers
Fri Jul 24 2026 • llm-stats.com
DeepSeek
DeepSeek-V4-Pro-Max
Input tokens$1.60
Output tokens$3.20
Best providerNovita
Alibaba Cloud / Qwen Team
Qwen3.5-397B-A17B
Input tokens$0.60
Output tokens$3.60
Best providerNovita
Notice missing or incorrect data?Start an Issue

Model Size

Parameter count comparison

1203.0B diff

DeepSeek-V4-Pro-Max has 1203.0B more parameters than Qwen3.5-397B-A17B, making it 303.0% larger.

DeepSeek
DeepSeek-V4-Pro-Max
1.6Tparameters
Alibaba Cloud / Qwen Team
Qwen3.5-397B-A17B
397.0Bparameters
1600.0B
DeepSeek-V4-Pro-Max
397.0B
Qwen3.5-397B-A17B

Context Window

Maximum input and output token capacity

DeepSeek-V4-Pro-Max accepts 1,048,576 input tokens compared to Qwen3.5-397B-A17B's 262,144 tokens. DeepSeek-V4-Pro-Max can generate longer responses up to 131,072 tokens, while Qwen3.5-397B-A17B is limited to 64,000 tokens.

DeepSeek
DeepSeek-V4-Pro-Max
Input1,048,576 tokens
Output131,072 tokens
Alibaba Cloud / Qwen Team
Qwen3.5-397B-A17B
Input262,144 tokens
Output64,000 tokens
Fri Jul 24 2026 • llm-stats.com

Input Capabilities

Supported data types and modalities

Qwen3.5-397B-A17B supports multimodal inputs, whereas DeepSeek-V4-Pro-Max does not.

Qwen3.5-397B-A17B can handle both text and other forms of data like images, making it suitable for multimodal applications.

DeepSeek-V4-Pro-Max

Text
Images
Audio
Video

Qwen3.5-397B-A17B

Text
Images
Audio
Video

License

Usage and distribution terms

DeepSeek-V4-Pro-Max is licensed under MIT, while Qwen3.5-397B-A17B uses Apache 2.0.

License differences may affect how you can use these models in commercial or open-source projects.

DeepSeek-V4-Pro-Max

MIT

Open weights

Qwen3.5-397B-A17B

Apache 2.0

Open weights

Release Timeline

When each model was launched

DeepSeek-V4-Pro-Max was released on 2026-04-23, while Qwen3.5-397B-A17B was released on 2026-02-16.

DeepSeek-V4-Pro-Max is 2 months newer than Qwen3.5-397B-A17B.

DeepSeek-V4-Pro-Max

Apr 23, 2026

3 months ago

2mo newer
Qwen3.5-397B-A17B

Feb 16, 2026

5 months ago

Knowledge Cutoff

When training data ends

Neither model specifies a knowledge cutoff date.

Unable to compare the recency of their training data.

No cutoff dates available

Provider Availability

DeepSeek-V4-Pro-Max is available from Novita, DeepInfra, DeepSeek, Fireworks, Together. Qwen3.5-397B-A17B is available from Novita.

DeepSeek-V4-Pro-Max

novita logo
Novita
Input Price:Input: $1.60/1MOutput Price:Output: $3.20/1M
deepinfra logo
Deepinfra
Input Price:Input: $1.74/1MOutput Price:Output: $3.48/1M
deepseek logo
DeepSeek
Input Price:Input: $1.74/1MOutput Price:Output: $3.48/1M
fireworks logo
Fireworks
Input Price:Input: $1.74/1MOutput Price:Output: $3.48/1M
together logo
Together
Input Price:Input: $1.74/1MOutput Price:Output: $3.48/1M

Qwen3.5-397B-A17B

novita logo
Novita
Input Price:Input: $0.60/1MOutput Price:Output: $3.60/1M
* Prices shown are per million tokens

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion

Key Takeaways

Larger context window (1,048,576 tokens)
Less expensive output tokens
Higher BrowseComp score (83.4% vs 69.0%)
Higher GDPval-AA score (44.4% vs 32.0%)
Higher GPQA score (90.1% vs 88.4%)
Higher Humanity's Last Exam score (48.2% vs 28.7%)
Higher IMO-AnswerBench score (89.8% vs 80.9%)
Higher SWE-bench Multilingual score (76.2% vs 69.3%)
Higher SWE-Bench Verified score (80.6% vs 76.4%)
Higher Terminal-Bench 2.0 score (67.9% vs 52.5%)
Higher Toolathlon score (51.8% vs 38.3%)
Alibaba Cloud / Qwen Team

Qwen3.5-397B-A17B

View details

Alibaba Cloud / Qwen Team

Supports multimodal inputs
Less expensive input tokens
Higher MMLU-Pro score (87.8% vs 87.5%)

Detailed Comparison

Interactive Arena

Judge for yourself.

Run your own prompts against DeepSeek-V4-Pro-Max and Qwen3.5-397B-A17B side-by-side, then vote on the output you prefer.

DeepSeek-V4-Pro-Max
✓ Preferred
Qwen3.5-397B-A17B
Open in Playground
AI Model Comparison Table
Feature
DeepSeek
DeepSeek-V4-Pro-Max
Alibaba Cloud / Qwen Team
Qwen3.5-397B-A17B

FAQ

Common questions about DeepSeek-V4-Pro-Max vs Qwen3.5-397B-A17B.

Which is better, DeepSeek-V4-Pro-Max or Qwen3.5-397B-A17B?

DeepSeek-V4-Pro-Max significantly outperforms across most benchmarks. DeepSeek-V4-Pro-Max is made by DeepSeek and Qwen3.5-397B-A17B is made by Alibaba Cloud / Qwen Team. The best choice depends on your use case — compare their benchmark scores, pricing, and capabilities above.

How does DeepSeek-V4-Pro-Max compare to Qwen3.5-397B-A17B in benchmarks?

DeepSeek-V4-Pro-Max scores CodeForces: 100.0%, HMMT Feb 26: 95.2%, LiveCodeBench: 93.5%, MathArena Apex: 90.2%, GPQA: 90.1%. Qwen3.5-397B-A17B scores MMLU-Redux: 94.9%, HMMT 2025: 94.8%, C-Eval: 93.0%, HMMT25: 92.7%, IFEval: 92.6%.

Is DeepSeek-V4-Pro-Max cheaper than Qwen3.5-397B-A17B?

Qwen3.5-397B-A17B is 2.7x cheaper for input tokens. DeepSeek-V4-Pro-Max costs $1.60/M input and $3.20/M output via novita. Qwen3.5-397B-A17B costs $0.60/M input and $3.60/M output via novita.

What are the context window sizes for DeepSeek-V4-Pro-Max and Qwen3.5-397B-A17B?

DeepSeek-V4-Pro-Max supports 1.0M tokens and Qwen3.5-397B-A17B supports 262K tokens. A larger context window lets you process longer documents, conversations, or codebases in a single request.

What are the main differences between DeepSeek-V4-Pro-Max and Qwen3.5-397B-A17B?

Key differences include context window (1.0M vs 262K), input pricing ($1.60 vs $0.60/M), multimodal support (no vs yes), licensing (MIT vs Apache 2.0). See the full comparison above for benchmark-by-benchmark results.

Who makes DeepSeek-V4-Pro-Max and Qwen3.5-397B-A17B?

DeepSeek-V4-Pro-Max is developed by DeepSeek and Qwen3.5-397B-A17B is developed by Alibaba Cloud / Qwen Team.