o1 vs DeepSeek V3: Benchmark Comparison
Detailed comparison of o1 and DeepSeek V3 covering benchmarks, pricing, context window, and compliance.
Key Specifications
| Specification | o1 | DeepSeek V3 |
|---|---|---|
| Vendor | openai | deepseek |
| Version | o1 | v3 |
| Release Date | 2024-12-17 | 2024-12-26 |
| Context Window | 200000 tokens | 64000 tokens |
| Input Modalities | text | text |
| Output Modalities | text | text |
| License | Proprietary | DeepSeek License |
| SOC2 | ✓ | ✗ |
| HIPAA | ✓ | ✗ |
| GDPR | ✓ | ✗ |
| ISO 27001 | ✓ | ✗ |
Benchmark Results
| Benchmark | o1 | DeepSeek V3 | Winner |
|---|---|---|---|
| ARC | 96.8 | — | o1 |
| BBH | 83.1 | 84.9 | DeepSeek V3 |
| GPQA | 57.5 | — | o1 |
| GSM8K | 88.9 | 89.3 | DeepSeek V3 |
| HUMANEVAL | 85.3 | 82.6 | o1 |
| IFEVAL | 82.2 | — | o1 |
| MATH | 55.7 | 61.6 | DeepSeek V3 |
| MMLU | 86.3 | 88.5 | DeepSeek V3 |
| MUSR | 71.8 | — | o1 |
| WINOGRANDE | 86.7 | — | o1 |
Pricing Comparison
| Tier (per Mtok) | o1 | DeepSeek V3 |
|---|---|---|
| Input | $15 | $0.27 |
| Output | $60 | $1.1 |
| Cache Read | $0 | $0.07 |
| Cache Write | $0 | $0.27 |
o1 対 DeepSeek V3
モデル概要
o1 and DeepSeek V3 are both notable options in the AI model market. This page compares their benchmarks, pricing, and compliance.
主要仕様
| ベンダー | リリース日 | コンテキストウィンドウ | ライセンス |
|---|---|---|---|
| Openai / Deepseek | 2024-12-17 / 2024-12-26 | 200K / 64K | Proprietary / DeepSeek License |
ベンチマークパフォーマンス
| ベンチマーク | o1 | DeepSeek V3 | 勝者 |
|---|---|---|---|
| ARC | 96.8 | — | A |
| BBH (BIG-Bench Hard) | 83.1 | 84.9 | B |
| GPQA | 57.5 | — | A |
| GSM8K (Grade School Math 8K) | 88.9 | 89.3 | Tie |
| HumanEval | 85.3 | 82.6 | A |
| IFEval | 82.2 | — | A |
| MATH | 55.7 | 61.6 | B |
| MMLU (Massive Multitask Language Understanding) | 86.3 | 88.5 | B |
| MUSR | 71.8 | — | A |
| WinoGrande | 86.7 | — | A |
料金比較
| 入力 | 出力 | キャッシュ読み取り | キャッシュ書き込み |
|---|---|---|---|
| — / — | — / — | — / — | — / — |
100万トークンあたり — A / B
強み & 弱み
o1
- ✅ MMLU score 86.3, strong knowledge reasoning.
- ✅ HumanEval 85.3, excellent code generation.
- ✅ GSM8K 88.9, robust math reasoning.
- ⚠️ 闭源专有模型，不支持自托管。
DeepSeek V3
- ✅ MMLU score 88.5, strong knowledge reasoning.
- ✅ HumanEval 82.6, excellent code generation.
- ✅ GSM8K 89.3, robust math reasoning.
- ✅ 采用 MoE 混合专家架构。
- ⚠️ 闭源专有模型，不支持自托管。
編集者コメント
o1 and DeepSeek V3 each have their strengths. Choose based on workload (code, long context, vision), referencing the tables above.
FAQ
Which model is better for coding tasks?
Refer to the HumanEval benchmark table; the model with a higher score is better suited for coding tasks.
Which model is cheaper?
Refer to the pricing comparison table above; the model with lower input/output prices is more cost-effective.
Which has a longer context window?
Refer to the key specifications table; the model with a larger context window is better for long documents.
参考文献
Editor's Take
See Editor's Take section.