Model Comparison

DeepSeek-V3.2 (Thinking) vs DeepSeek-V3

DeepSeek-V3.2 (Thinking) significantly outperforms across most benchmarks. DeepSeek-V3.2 (Thinking) is 1.5x cheaper per token.

Want to compare interactively?Try the playground

Performance Benchmarks

Comparative analysis across standard metrics

4 benchmarks

DeepSeek-V3.2 (Thinking) outperforms in 4 benchmarks (GPQA, LiveCodeBench, MMLU-Pro, SWE-Bench Verified), while DeepSeek-V3 is better at 0 benchmarks.

DeepSeek-V3.2 (Thinking) significantly outperforms across most benchmarks.

Tue May 26 2026 • llm-stats.com

Arena Performance

Human preference votes

Pricing Analysis

Price comparison per million tokens

DeepSeek-V3.2 (Thinking) costs less

For input processing, DeepSeek-V3.2 (Thinking) ($0.28/1M tokens) is 1.0x more expensive than DeepSeek-V3 ($0.27/1M tokens).

For output processing, DeepSeek-V3.2 (Thinking) ($0.42/1M tokens) is 2.6x cheaper than DeepSeek-V3 ($1.10/1M tokens).

In conclusion, DeepSeek-V3 is more expensive than DeepSeek-V3.2 (Thinking).*

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

Lowest available price from all providers

Tue May 26 2026 • llm-stats.com

DeepSeek-V3.2 (Thinking)

Input tokens$0.28

Output tokens$0.42

Best providerDeepSeek

DeepSeek-V3

Input tokens$0.27

Output tokens$1.10

Best providerDeepSeek

Notice missing or incorrect data?Start an Issue→

Model Size

Parameter count comparison

14.0B diff

DeepSeek-V3.2 (Thinking) has 14.0B more parameters than DeepSeek-V3, making it 2.1% larger.

DeepSeek-V3.2 (Thinking)

685.0Bparameters

DeepSeek-V3

671.0Bparameters

685.0B

DeepSeek-V3.2 (Thinking)

671.0B

DeepSeek-V3

Context Window

Maximum input and output token capacity

Both models have the same input context window of 131,072 tokens. DeepSeek-V3 can generate longer responses up to 131,072 tokens, while DeepSeek-V3.2 (Thinking) is limited to 65,536 tokens.

DeepSeek-V3.2 (Thinking)

Input131,072 tokens

Output65,536 tokens

DeepSeek-V3

Input131,072 tokens

Output131,072 tokens

Tue May 26 2026 • llm-stats.com

License

Usage and distribution terms

DeepSeek-V3.2 (Thinking) is licensed under MIT, while DeepSeek-V3 uses MIT + Model License (Commercial use allowed).

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

DeepSeek-V3.2 (Thinking)

MIT

Open weights

DeepSeek-V3

MIT + Model License (Commercial use allowed)

Open weights

Release Timeline

When each model was launched

DeepSeek-V3.2 (Thinking) was released on 2025-12-01, while DeepSeek-V3 was released on 2024-12-25.

DeepSeek-V3.2 (Thinking) is 11 months newer than DeepSeek-V3.

DeepSeek-V3.2 (Thinking)

Dec 1, 2025

5 months ago

11mo newer

DeepSeek-V3

Dec 25, 2024

1.4 years 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-V3.2 (Thinking) is available from DeepSeek. DeepSeek-V3 is available from DeepSeek.

DeepSeek-V3.2 (Thinking)

DeepSeek

Input Price:Input: $0.28/1MOutput Price:Output: $0.42/1M

DeepSeek-V3

DeepSeek

Input Price:Input: $0.27/1MOutput Price:Output: $1.10/1M

* Prices shown are per million tokens

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion→

Key Takeaways

DeepSeek-V3.2 (Thinking)

View details

DeepSeek

Less expensive output tokens

Higher GPQA score (82.4% vs 59.1%)

Higher LiveCodeBench score (83.3% vs 37.6%)

Higher MMLU-Pro score (85.0% vs 75.9%)

Higher SWE-Bench Verified score (73.1% vs 42.0%)

DeepSeek-V3

View details

DeepSeek

Less expensive input tokens

Detailed Comparison

AI Model Comparison Table
Feature	DeepSeek-V3.2 (Thinking)	DeepSeek-V3

FAQ

Common questions about DeepSeek-V3.2 (Thinking) vs DeepSeek-V3.

Which is better, DeepSeek-V3.2 (Thinking) or DeepSeek-V3?

DeepSeek-V3.2 (Thinking) significantly outperforms across most benchmarks. DeepSeek-V3.2 (Thinking) is made by DeepSeek and DeepSeek-V3 is made by DeepSeek. The best choice depends on your use case — compare their benchmark scores, pricing, and capabilities above.

How does DeepSeek-V3.2 (Thinking) compare to DeepSeek-V3 in benchmarks?

DeepSeek-V3.2 (Thinking) scores AIME 2025: 93.1%, HMMT 2025: 90.2%, MMLU-Pro: 85.0%, LiveCodeBench: 83.3%, GPQA: 82.4%. DeepSeek-V3 scores DROP: 91.6%, CLUEWSC: 90.9%, MATH-500: 90.2%, MMLU-Redux: 89.1%, MMLU: 88.5%.

Is DeepSeek-V3.2 (Thinking) cheaper than DeepSeek-V3?

DeepSeek-V3 is 1.0x cheaper for input tokens. DeepSeek-V3.2 (Thinking) costs $0.28/M input and $0.42/M output via deepseek. DeepSeek-V3 costs $0.27/M input and $1.10/M output via deepseek.

What are the context window sizes for DeepSeek-V3.2 (Thinking) and DeepSeek-V3?

DeepSeek-V3.2 (Thinking) supports 131K tokens and DeepSeek-V3 supports 131K tokens. A larger context window lets you process longer documents, conversations, or codebases in a single request.

What are the main differences between DeepSeek-V3.2 (Thinking) and DeepSeek-V3?

Key differences include input pricing ($0.28 vs $0.27/M), licensing (MIT vs MIT + Model License (Commercial use allowed)). See the full comparison above for benchmark-by-benchmark results.