DeepSeek V3.2 vs GPT-4o
Compare pricing, context limits, capabilities, and benchmarks side by side.
Presets:DeepSeek V3.2 vs DeepSeek V4 FlashGPT-4o vs Claude 3.5 SonnetDeepSeek V3.2 vs GPT-4oGemini 1.5 Pro vs GPT-4oClaude 3.5 Haiku vs GPT-4o Mini
Selected Candidates:
DeepSeek V3.2
deepseek-ai • $0.25/1M
GPT-4o
openai • $2.50/1M
| Specification | DeepSeek V3.2 | GPT-4o |
|---|---|---|
Input Price Per 1M prompt tokens | $0.25Lowest | $2.50 |
Output Price Per 1M completion tokens | $0.39Lowest | $10.00 |
Context Limit Max tokens | 164k tokensLargest | 128k tokens |
Error Rate Hallucination % | 670.0% | 150.0%Lowest |
| Provider | deepseek-ai | openai |
| Vision Support | No | Yes |
| Function Calling | Yes | Yes |
| Reasoning Mode | Standard | Standard |
Benchmark Scores
Standardized intelligence, coding, and mathematical evaluations.
| Benchmark | DeepSeek V3.2 | GPT-4o |
|---|---|---|
Intelligence Index Overall AI reasoning, coding, and knowledge score. | 32.1Top | 18.6 |
Coding Index Code generation, debugging, and software engineering. | 34.6Top | 16.6 |
Math Index Mathematical problem solving capability. | 59Top | 6 |
MMLU Pro Multi-task understanding across 14 domains. | 0.837Top | 0.748 |
GPQA Graduate-level reasoning questions. | 0.751Top | 0.521 |
LiveCodeBench Competitive programming problems. | 0.593Top | 0.317 |
MATH-500 Contest and high-school math problems. | N/A | 0.795 |
AIME 2025 Mathematics Examination 2025. | 0.59Top | 0.06 |
Humanity's Last Exam Frontier multi-disciplinary benchmark. | 0.105Top | 0.029 |
DeepSeek V3.2
GPT-4o
