Requesty

deepseek-v4-pro vs grok-4

Side-by-side comparison of deepseek-v4-pro and grok-4: benchmarks, pricing, context window and capabilities. Both are accessible through Requesty's unified API. deepseek-v4-pro outperforms grok-4 on 7 of 7 shared benchmarks.

Benchmark comparison

Intelligence Indexreasoning
deepseek-v4-pro51.5%
grok-441.5%
Coding Indexcoding
deepseek-v4-pro47.5%
grok-440.5%
Math Indexmath
deepseek-v4-proN/A
grok-492.7%
GPQA Diamondreasoning
deepseek-v4-pro88.8%
grok-487.7%
AIME 2025math
deepseek-v4-proN/A
grok-492.7%
LiveCodeBenchcoding
deepseek-v4-proN/A
grok-481.9%
Terminal-Bench Hardagentic
deepseek-v4-pro46.2%
grok-437.9%
τ²-Benchagentic
deepseek-v4-pro96.2%
grok-474.9%
SciCodecoding
deepseek-v4-pro50.0%
grok-445.7%
MMLU Proknowledge
deepseek-v4-proN/A
grok-486.6%
Humanity's Last Examreasoning
deepseek-v4-pro35.9%
grok-423.9%

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-progrok-4
Input price / 1M$0.43$3.00
Output price / 1M$0.87$15.00
Context window1M tokens256K tokens
Max output384K tokensN/A
Vision inputN/AYes
Tool callingYesYes
ReasoningN/AN/A
Prompt cachingYesYes
Computer useN/AYes
ProviderDeepSeekxAI Corp.

Questions people ask

Is deepseek-v4-pro better than grok-4?
deepseek-v4-pro outperforms grok-4 on 7 of 7 shared benchmarks. See the benchmark comparison above for specifics: deepseek-v4-pro and grok-4 have different strengths across reasoning, coding, math and multimodal tasks.
Which is cheaper, deepseek-v4-pro or grok-4?
deepseek-v4-pro is cheaper. deepseek-v4-pro costs $0.43/$0.87 per 1M input/output tokens, while grok-4 costs $3.00/$15.00.
Can I use deepseek-v4-pro and grok-4 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" and "xai/grok-4", no other code changes needed.
What are the context windows?
deepseek-v4-pro supports up to 1M tokens of context. grok-4 supports up to 256K 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 and grok-4 with one line of code

Requesty provides a single OpenAI-compatible API for 400+ models. Change the model parameter, not your code.

Get started free