DeepSeek V3.2 vs GPT-4.1
Head to head on 15 shared benchmarks, with price, context window, and release dates.
Benchmark scores
56evals · sort by any column, ⤢ to expand fullscreen
| Model | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DeepSeek V3.2 DeepSeek | 74.2% | 59.0% | 47.9% | 1511elo | 22.1% | 75.1% | 11.2% | 49.0% | 59.3% | 83.7% | 38.7% | 70.0% | 68.2% | 32.6% | 78.9% | 74.7% | - | - | - | - | - | - | 0.0% | - | 16.0% | - | 66.8 | - | 31.3% | 68.0% | 65.9% | 64.6% | 50.0% | 48.2% | 70.4% | 85.0% | 77.2% | - | - | - | - | 57.5% | - | - | - | - | - | 8.0% | - | 59.0% | - | - | 65 | 5.1% | - | - | 51.7% |
| GPT-4.1 OpenAI | 52.4% | 34.7% | 63.1% | 1417elo | 5.5% | 66.6% | 4.2% | 43.0% | 45.7% | 80.6% | 38.1% | 39.6% | 75.1% | 13.6% | 47.1% | - | 0.0% | 43.7% | 26.7% | 61.5% | 1.64 | 83.7% | - | 1.55 | - | 12.4% | - | 60.9% | - | - | - | - | - | - | - | - | - | 70.8% | 0.49 | 0.49 | 91.3% | - | 32.1% | 32.1% | 69.4% | 50.5% | 65.9% | - | 8.57 | - | 31.1% | 10.0% | - | - | 92.0% | 92.0% | 48.0% |
2 / 2 models
| Attribute | DeepSeek V3.2 DeepSeek | GPT-4.1 OpenAI |
|---|---|---|
| Description | DeepSeek V3.2 is an AI model from DeepSeek, released with open weights. | GPT-4.1 is an AI model from OpenAI. |
| Family | Deepseek | GPT |
| Params | - | - |
| Open weights | open | closed |
| License | mit | proprietary |
| Released | 1 Dec 2025 | 14 Apr 2025 |
| Context | 163,840 | 1,047,576 |
| Input $/1M | $0.28 | $2.00 |
| Output $/1M | $0.42 | $8.00 |
| Cache read $/1M | $0.130 | $0.500 |
| Cache write $/1M | - | - |
| Speed | 0 tok/s | 0 tok/s |
| Latency | 0.00 s | 0.00 s |
| Modalities | text | image, text, file |
| Providers | - | - |
| Publisher | DeepSeek | OpenAI |
| Reported scores | Aider Polyglot Benchmark74.2% (Aider) MathArena57.47% GPQA Diamond75.1% (AA) Humanity's Last Exam (HLE)11.2% (AA) | ScholarSearch8.57 Social Sciences & Humanities (%) GPQA Diamond66.6% (AA) Humanity's Last Exam (HLE)4.2% (AA) IFBench43.0% (AA) |
FAQ
- Is DeepSeek V3.2 or GPT-4.1 better?
- Across the 15 benchmarks both models report on Sophon, DeepSeek V3.2 leads on 13 of them. Which one is "better" depends on the benchmark - the table above shows every shared score.
- How do DeepSeek V3.2 and GPT-4.1 compare on GPQA Diamond?
- DeepSeek V3.2 scores 75.1% and GPT-4.1 scores 66.6% on GPQA Diamond.
- Which is cheaper, DeepSeek V3.2 or GPT-4.1?
- DeepSeek V3.2 is cheaper at $0.42 per million output tokens against $8.00 for GPT-4.1 - about 19.0x.
- When were DeepSeek V3.2 and GPT-4.1 released?
- DeepSeek V3.2 was released 1 Dec 2025 by DeepSeek. GPT-4.1 was released 14 Apr 2025 by OpenAI.
Comparing something else? Build your own side-by-side.