deepseek-v4-pro-0813 vs deepseek-v3.2
Side-by-side comparison of deepseek-v4-pro-0813 and deepseek-v3.2: benchmarks, pricing, context window and capabilities. Both are accessible through Requesty's unified API. deepseek-v4-pro-0813 outperforms deepseek-v3.2 on 4 of 4 shared benchmarks.
deepseek-v4-pro-0813
Input / 1M
$1.32
Output / 1M
$3.96
Context
1M
Model ID
deepseek/deepseek-v4-pro-0813
deepseek-v3.2
Input / 1M
$0.56
Output / 1M
$1.68
Context
164K
Model ID
vertex/deepseek-v3.2
Benchmark comparison
Intelligence Indexreasoning
deepseek-v4-pro-081342.1%
deepseek-v3.225.3%
Coding Indexcoding
deepseek-v4-pro-081368.8%
deepseek-v3.244.2%
Math Indexmath
deepseek-v4-pro-0813N/A
deepseek-v3.292.0%
GPQA Diamondreasoning
deepseek-v4-pro-081392.8%
deepseek-v3.284.0%
AIME 2025math
deepseek-v4-pro-0813N/A
deepseek-v3.292.0%
LiveCodeBenchcoding
deepseek-v4-pro-0813N/A
deepseek-v3.286.2%
Terminal-Bench Hardagentic
deepseek-v4-pro-0813N/A
deepseek-v3.235.6%
τ²-Benchagentic
deepseek-v4-pro-0813N/A
deepseek-v3.290.6%
SciCodecoding
deepseek-v4-pro-081351.0%
deepseek-v3.2N/A
MMLU Proknowledge
deepseek-v4-pro-0813N/A
deepseek-v3.286.2%
Humanity's Last Examreasoning
deepseek-v4-pro-081341.0%
deepseek-v3.224.6%
Scores sourced from official model cards, Artificial Analysis, and public leaderboards. Benchmarks measure specific skills and don't capture every aspect of model quality.
Pricing & specifications
| deepseek-v4-pro-0813 | deepseek-v3.2 | |
|---|---|---|
| Input price / 1M | $1.32 | $0.56 |
| Output price / 1M | $3.96 | $1.68 |
| Context window | 1M tokens | 164K tokens |
| Max output | 384K tokens | 66K tokens |
| Vision input | N/A | N/A |
| Tool calling | Yes | Yes |
| Reasoning | Yes | Yes |
| Prompt caching | Yes | Yes |
| Computer use | N/A | N/A |
| Provider | DeepSeek | Google LLC (Vertex AI) |
Questions people ask
Is deepseek-v4-pro-0813 better than deepseek-v3.2?
deepseek-v4-pro-0813 outperforms deepseek-v3.2 on 4 of 4 shared benchmarks. See the benchmark comparison above for specifics: deepseek-v4-pro-0813 and deepseek-v3.2 have different strengths across reasoning, coding, math and multimodal tasks.
Which is cheaper, deepseek-v4-pro-0813 or deepseek-v3.2?
deepseek-v3.2 is cheaper. deepseek-v4-pro-0813 costs $1.32/$3.96 per 1M input/output tokens, while deepseek-v3.2 costs $0.56/$1.68.
Can I use deepseek-v4-pro-0813 and deepseek-v3.2 through the same API?
Yes. Requesty provides a single OpenAI-compatible API that routes to both. Change just the "model" parameter to switch between "deepseek/deepseek-v4-pro-0813" and "vertex/deepseek-v3.2", no other code changes needed.
What are the context windows?
deepseek-v4-pro-0813 supports up to 1M tokens of context. deepseek-v3.2 supports up to 164K tokens. Longer context means you can feed larger documents or codebases in a single prompt, though quality often degrades past 128K for most models.
Switch between deepseek-v4-pro-0813 and deepseek-v3.2 with one line of code
Requesty provides a single OpenAI-compatible API for 600+ models. Change the model parameter, not your code.