DeepSeek V4 Pro 0424 vs GPT-6.1 Sol
完整资料:DeepSeek V4 Pro 0424(DeepSeek)和 GPT-6.1 Sol(OpenAI),包括全部基准测试、价格和排名。
更新于 · 数据来自 14 个来源
截至 2026年11月16日,两者共同的基准测试太少,无法判断质量高下。DeepSeek V4 Pro 0424 更便宜,对方的价格是它的 1.9 倍(每百万 token 混合价格 $2.09 比 $4.00)。在两者都有成绩的所有类别中,GPT-6.1 Sol 均领先。
要点(截至 2026年11月16日)
- DeepSeek V4 Pro 0424 每百万 token 混合价格为 $2.09(标价,2026年11月16日)。
- GPT-6.1 Sol 每百万 token 混合价格为 $4.00(标价,2026年11月16日)。
- 在两者都有结果的 5 项基准测试中,GPT-6.1 Sol 领先 5 项(BenchLeader,2026年11月16日)。
按用途怎么选
推荐模型的数值列在前面。类别得分采用 BenchLeader 标尺(50 为模型平均水平,15 分为一个标准差)。领先半个标准差及以上为“高”置信度;相差不到 1 分则视为“均可”。
主要差异
- 共同的基准测试太少,无法给出结论。
- GPT-6.1 Sol (max) 在智能体与工具、长上下文方面更强。
- DeepSeek V4 Pro 0424 (high) 更便宜,对方的价格是它的 1.9 倍(每百万 token 混合价格 $2.09 比 $4.00)。
概览
| 规格 | DeepSeek V4 Pro 0424 | GPT-6.1 Sol |
|---|---|---|
| 提供商 | DeepSeek | OpenAI |
| 发布日期 | 2026年4月24日 | 2026年9月29日 |
| 上下文窗口(token) | 1,000,000 | 1,050,000 |
| 对比时的配置 | DeepSeek V4 Pro 0424 (high) | GPT-6.1 Sol (max) |
| 开放权重 | 是 | 否 |
| 每百万 token 价格(输入 / 输出 / 缓存输入) | $1.60 / $3.20 / – | $2.00 / $10.00 / $0.100 |
逐类别对比
类别得分采用 BenchLeader 标尺:50 为参评模型的平均水平,15 分为一个标准差。差距不到 1 分视为持平。
| 类别 | DeepSeek V4 Pro 0424 (high) | GPT-6.1 Sol (max) | 领先者 |
|---|---|---|---|
| 推理 | – | 71.4 | 无法比较 |
| 编程 | – | 71.1 | 无法比较 |
| 智能体与工具 | 59.7 | 64.3 | GPT-6.1 Sol |
| 数学 | – | 71.6 | 无法比较 |
| 知识 | – | 71.5 | 无法比较 |
| 指令遵循 | 70.0 | – | 无法比较 |
| 人类偏好 | – | 67.1 | 无法比较 |
| 多模态 | – | 65.6 | 无法比较 |
| 长上下文 | 59.6 | 66.2 | GPT-6.1 Sol |
| 综合 | – | 67.9 | 无法比较 |
同一图表对比
各领域采用与 Index 相同的标尺,平均模型为 50 分。单击某个领域可查看构成它的基准测试。 可从列表中最多选择 5 个模型。
虚线:该模型在所经过的轴上没有结果,线条只连接相邻的点。
各来源的分歧
在两者都有结果的 5 项基准测试中,DeepSeek V4 Pro 0424 领先 0 项,GPT-6.1 Sol 领先 5 项。
各发布方使用各自的任务、评测框架和推理强度设置,因此总体结论可能不同;BenchLeader 将它们统一到同一标尺上。† 表示同一模型在另一推理强度设置下的测量值(发布方未测试本页对比的设置)。
价格、速度与延迟
标价按每 1 个输出 token 对应 3 个输入 token 混合计算。每项任务的价格采用方法论页面中的 token 数量。每项基准测试任务的成本、token 数和耗时均为实测值,取自 Artificial Analysis 整个 Intelligence Index 测试的平均值。
| 指标 | DeepSeek V4 Pro 0424 (high) | GPT-6.1 Sol (max) |
|---|---|---|
| 每百万 token 混合价格 | $2.00 | $4.00 |
| 每百万 token 输入 / 输出 | $1.60 / $3.20 | $2.00 / $10.00 |
| 每百万 token 缓存输入 | – | $0.100 |
| 每项任务价格:聊天回复 | $0.0016 | $0.0038 |
| 每项任务价格:总结一份 30 页的报告 | $0.021 | $0.030 |
| 每项任务价格:代码编辑 | $0.014 | $0.027 |
| 每项任务价格:智能体编程会话 | $0.109 | $0.160 |
| 每项任务价格:结构化信息提取 | $0.0038 | $0.0060 |
| 智能体编程会话(输入大多命中缓存) | – | $0.074 |
| 每项基准测试任务成本(AA Index) | – | $1.27 |
| 每项任务输出 token(AA Index) | – | 68k |
| 其中推理部分 | – | 49k |
| 每项任务耗时(AA Index) | – | 23 min |
| 输出速度 | – | 50 tok/s |
| 首次回答时间 | – | 166.7 秒 |
| 完整回答耗时(最近一次测量 2026年11月13日) | – | 429.6 秒 |
每个模型都按其排名最高的推理强度设置进行对比。DeepSeek V4 Pro 0424 (high)。GPT-6.1 Sol (max) 得分 66.6,首个回答 token 需 167 秒;GPT-6.1 Sol (xhigh) 为 63.6(68 秒),GPT-6.1 Sol (high) 为 63.4(33 秒)。推理强度越高,回答前思考越久:得分更高,但等待更长、计费 token 也更多。 推理强度如何影响得分和成本
全部基准测试
共 159 项基准测试,按类别分组;点击名称可查看该排行榜及领先模型。† 表示同一模型在另一推理强度设置下的测量值,仅供参考,不计入胜场。
| 指标 | 发布方 | DeepSeek V4 Pro 0424 (high) | GPT-6.1 Sol (max) |
|---|---|---|---|
| BenchLeader Index | BenchLeader | – | 66.6 |
| 智能体与工具得分 | BenchLeader | 59.7 | 64.3 |
| 指令遵循得分 | BenchLeader | 70.0 | – |
| 长上下文得分 | BenchLeader | 59.6 | 66.2 |
| 编程得分 | BenchLeader | – | 71.1 |
| 综合得分 | BenchLeader | – | 67.9 |
| 人类偏好得分 | BenchLeader | – | 67.1 |
| 知识得分 | BenchLeader | – | 71.5 |
| 数学得分 | BenchLeader | – | 71.6 |
| 多模态得分 | BenchLeader | – | 65.6 |
| 推理得分 | BenchLeader | – | 71.4 |
| 混合价格(美元/百万 token) | BenchLeader | $2.00 | $4.00 |
| 输出速度 | BenchLeader | – | 50 tok/s |
| 首次回答时间 | BenchLeader | – | 166.7 秒 |
| 上下文窗口 | BenchLeader | 1,000,000 token | 1,050,000 token |
| 推理 | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | – | 95.4% |
| SimpleBench | SimpleBench | – | 82.9%† |
| Arena Hard Prompts | Arena (LMArena) | – | 1505 |
| LiveBench 2.0 Reasoning | LiveBench | – | 92.6% |
| LiveBench 1.0 Reasoning | LiveBench | – | 92.6% |
| GPQA Diamond (AA) | Artificial Analysis | 90.5% | – |
| Humanity's Last Exam (AA) | Artificial Analysis | 35.2% | 52.9% |
| HLE-Diamond (AA) | Artificial Analysis | – | 58.8% |
| ARC-AGI-1 | ARC Prize | – | 96.5% |
| ARC-AGI-2 | ARC Prize | – | 94.2% |
| ARC-AGI-3 | ARC Prize | – | 96.2% |
| CritPt | Artificial Analysis | 10.0% | 31.7% |
| CritPt v2 hard subset (AA) | Artificial Analysis | – | 69.5% |
| MysteryMechanism | Vals AI | – | 46.4% |
| Arena Expert | Arena (LMArena) | – | 1525 |
| Chess Puzzles | Epoch AI Benchmarking Hub | – | 68.0% |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | – | 80.0% |
| LMCA | Epoch AI Benchmarking Hub | – | 62.8% |
| DTBench | Epoch AI Benchmarking Hub | – | 96.5% |
| 编程 | |||
| SciCode | SciCode | – | 54.2% |
| FrontierCode | Cognition | – | 50.2%† |
| Arena Coding | Arena (LMArena) | – | 1540 |
| Arena WebDev | Arena (LMArena) | – | 1743 |
| LiveBench 2.0 Coding | LiveBench | – | 82.1% |
| LiveBench 1.0 Coding | LiveBench | – | 80.4% |
显示全部 159 项基准测试
| 指标 | 发布方 | DeepSeek V4 Pro 0424 (high) | GPT-6.1 Sol (max) |
|---|---|---|---|
| 编程 | |||
| SciCode (AA) | Artificial Analysis | 50.8%† | 54.2% |
| IOI | Vals AI | – | 96.9% |
| Code Migration | Vals AI | – | 65.1% |
| Code Migration v1.1 | Vals AI | – | 77.3% |
| ProgramBench | Vals AI | – | 3.0% |
| SWE-bench Multimodal | Vals AI | – | 43.5% |
| Vibe Code Bench v1.1 | Vals AI | – | 88.9% |
| SWE Atlas v1.1 | Scale AI SEAL | – | 65.9%† |
| SWE Atlas v1.1: Codebase QnA | Scale AI SEAL | – | 71.0%† |
| SWE Atlas v1.1: Test Writing | Scale AI SEAL | – | 66.7%† |
| SWE Atlas v1.1: Refactoring | Scale AI SEAL | – | 60.0%† |
| SWE-Bench Pro V2 | Scale AI SEAL | – | 96.1%† |
| MirrorCode | Epoch AI Benchmarking Hub | – | 61.1% |
| ALE-Bench | Epoch AI Benchmarking Hub | – | 2743.8 |
| WeirdML v3 | WeirdML | – | 38.3%† |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 14.7%† | 56.1% |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | 64.8% | – |
| Startrise Frontend | Startrise | – | 78.6%† |
| GameBench 2 Overall | DuelLab GameBench 2 | – | 84.9† |
| GameBench 2 by reasoning setting | DuelLab GameBench 2 | – | 92.0† |
| Design Arena Website | Design Arena | – | 1315† |
| FrontierCode 1.1 (AA) | Artificial Analysis | – | 49.1% |
| FrontierCode 2.0 (AA) | Artificial Analysis | – | 40.6% |
| 智能体与工具 | |||
| Terminal-Bench | Terminal-Bench | – | 60.9% |
| APEX-Agents | Mercor | – | 60.0% |
| Arena Agent | Arena (LMArena) | – | 12.5 |
| LiveBench Long Agentic | LiveBench | – | 65.6% |
| LiveBench Browser | LiveBench | – | 93.3% |
| LiveBench Agentic Coding | LiveBench | – | 54.5% |
| Terminal-Bench Hard | Artificial Analysis | 41.7% | – |
| τ²-Bench Telecom (AA) | Artificial Analysis | 94.2% | – |
| HiL-Bench | Scale AI SEAL | – | 62.3%† |
| GDPval-AA v2.1, normalised score | Artificial Analysis | – | 54.8% |
| GDPval-AA v2.1 | Artificial Analysis | 1160 | 1607 |
| τ³-Banking (AA) | Artificial Analysis | 26.2% | – |
| ITBench SRE (AA) | Artificial Analysis | 38.3%† | – |
| Analyst Agent (AA) | Artificial Analysis | 18.8%† | 50.0% |
| APEX-Agents (AA) | Artificial Analysis | 24.3%† | – |
| CyberBench | Vals AI | – | 39.3% |
| Finance Agent v2 | Vals AI | – | 52.0% |
| Finance Agent v2.1 | Vals AI | – | 61.8% |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | – | 6.3% |
| Harvey's Legal Agent Benchmark v1 | Vals AI | – | 5.4% |
| Legal Research Bench | Vals AI | – | 38.5% |
| Robo-bench | Vals AI | – | 28.3% |
| SREBench | Vals AI | – | 50.8% |
| Tax Agent Bench | Vals AI | – | 21.8% |
| Terminal-Bench 4.0 (Vals) | Vals AI | – | 55.0% |
| Terminal-Bench Science | Vals AI | – | 52.9% |
| Arena Search | Arena (LMArena) | – | 1249 |
| Arena Agent Task Completion | Arena (LMArena) | – | 11.7 |
| PostTrainBench | Epoch AI Benchmarking Hub | – | 38.8% |
| Vending-Bench 2 | Andon Labs | – | 12860.6† |
| SpireClimb | Epoch AI Benchmarking Hub | – | 23.8% |
| AutomationBench | Artificial Analysis | 56.3%† | 64.9% |
| GDP.pdf (AA) | Artificial Analysis | 13.4%† | 31.0% |
| Harvey LAB-AA v1.1 | Harvey | – | 6.9% |
| EnterpriseOps-Gym | Artificial Analysis | 40.4%† | – |
| AA-Briefcase v1.1 | Artificial Analysis | 922† | 1555 |
| GDPx | GDPx by SemiAnalysis | – | 8.4% |
| Design Arena Agentic | Design Arena | – | 1375 |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | – | 58.1% |
| MKA (AA) | Artificial Analysis | – | 47.7% |
| FACTS Search | – | 92.8%† | |
| 数学 | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | – | 93.7% |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | – | 100.0% |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | – | 100.0% |
| ProofBench | Vals AI | – | 99.0%† |
| LiveBench Mathematics | LiveBench | – | 96.8% |
| ArXivMath | MathArena | – | 92.8% |
| BrokenArXiv | MathArena | – | 83.1% |
| Arena Maths | Arena (LMArena) | – | 1499 |
| FrontierMath Erdős | Epoch AI Benchmarking Hub | – | 4.4% |
| 知识 | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | – | 73.9% |
| LiveBench 2.0 Data Analysis | LiveBench | – | 82.7% |
| LiveBench Research | LiveBench | – | 90.0% |
| LiveBench 1.0 Data Analysis | LiveBench | – | 82.7% |
| AA-Omniscience v1.1 | Artificial Analysis | – | 42.0 |
| PRBench Finance | Scale AI SEAL | – | 46.2%† |
| PRBench Legal | Scale AI SEAL | – | 47.9%† |
| BioMysteryBench | Vals AI | – | 79.6% |
| ClaimBench | Vals AI | – | 100.0% |
| Excel Modeling Benchmark | Vals AI | – | 70.8% |
| Excel Modeling Benchmark v1.1 | Vals AI | – | 71.1% |
| MedCode | Vals AI | – | 48.8% |
| MedScribe | Vals AI | – | 86.5% |
| Public Benefits Bench v1.1 | Vals AI | – | 59.3% |
| Access to Justice Benchmark | Vals AI | – | 65.8% |
| Financial Coaching Bench | Vals AI | – | 68.5% |
| Immigration Bench | Vals AI | – | 67.8% |
| EBR-bench | Epoch AI Benchmarking Hub | – | 54.3% |
| GDP.pdf | Epoch AI Benchmarking Hub | – | 35.0% |
| AA-Omniscience v1.1: accuracy | Artificial Analysis | – | 61.8% |
| AA-Omniscience v1.1: non-hallucination | Artificial Analysis | – | 48.1% |
| HealthBench Professional (AA) | OpenAI | – | 65.1% |
| FACTS Parametric | – | 83.8%† | |
| 指令遵循 | |||
| LiveBench Language | LiveBench | – | 90.1% |
| LiveBench Instruction Following | LiveBench | – | 74.2% |
| IFBench | Artificial Analysis | 71.3% | – |
| MultiChallenge | Scale AI SEAL | – | 82.0%† |
| Arena Instruction Following | Arena (LMArena) | – | 1482 |
| 人类偏好 | |||
| Arena Text | Arena (LMArena) | – | 1481 |
| EQ-Bench 4 | EQ-Bench | – | 1272† |
| Arena Creative Writing | Arena (LMArena) | – | 1449 |
| Arena Multi-turn | Arena (LMArena) | – | 1486 |
| Arena Longer Queries | Arena (LMArena) | – | 1488 |
| 多模态 | |||
| Arena Vision | Arena (LMArena) | – | 1286 |
| LiveBench Design | LiveBench | – | 79.6% |
| MMMU-Pro | Artificial Analysis | – | 86.0% |
| SAGE | Vals AI | – | 90.5% |
| Humanity's Sixth Sense | Scale AI SEAL | – | 46.6% |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | – | 43.2%† |
| Furniture Assembly | Epoch AI Benchmarking Hub | – | 80.0% |
| VIALS-AA | Handshake AI | – | 59.0% |
| Code2Video-AA | Artificial Analysis | – | 1369 |
| AA-3D | Artificial Analysis | – | 1121 |
| FACTS Multimodal | – | 52.6%† | |
| 长上下文 | |||
| AA-LCR v1.1 | Artificial Analysis | 70.3% | 83.0% |
| MLCR | Artificial Analysis | 9.4%† | 38.3% |
| FACTS Grounding | – | 78.0%† | |
| 综合 | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | – | 168.5† |
| LiveBench 2.0 | LiveBench | – | 81.4% |
| LiveBench 1.0 | LiveBench | – | 81.6% |
| AA Intelligence Index v5.0.1 | Artificial Analysis | – | 54.6 |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 30.1 | 51.8 |
| AA Progress Index v1 | Artificial Analysis | 287.7 | 348.5 |
| Vals Index | Vals AI | – | 63.9 |
| Vals Finance Index | Vals AI | – | 55.8% |
| Vals RSI Index | Vals AI | – | 26.6% |
| Vals Science Index | Vals AI | – | 59.6% |
| Vals Router Index | Vals AI | – | 63.9%† |
| Vals Legal Index | Vals AI | – | 68.7% |
| 安全与诚实 | |||
| DistressBench | Scale AI SEAL | – | 74.2%† |
| 元数据 | |||
| AA Openness Index | Artificial Analysis | 50 | – |
数据截至 2026年11月16日。对比的是每个模型的最佳配置;每项分数的来源链接都在模型页面上。
DuelLab GameBench 2, release gb2-20261111-7b598bb13de4 (https://benchmarks.duellab.org/releases/gb2-20261111-7b598bb13de4/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.
变化
- GPT-6.1 Sol (max):8 项新结果(LiveBench 2.0, LiveBench Long Agentic, LiveBench 2.0 Reasoning…);1 项分数修订。
- DeepSeek V4 Pro 0424 (high):1 项新结果(GDPval-AA v2.1)。
- GPT-6.1 Sol (max):1 项分数修订。
- GPT-6.1 Sol (max):1 项分数修订。
- GPT-6.1 Sol (max):1 项新结果(PostTrainBench)。
- GPT-6.1 Sol (max):1 项新结果(MirrorCode);1 项分数修订。
- GPT-6.1 Sol (max):BenchLeader Index 67.4 → 66.9。
- GPT-6.1 Sol (max):2 项分数修订。
- GPT-6.1 Sol (max):3 项分数修订。
- GPT-6.1 Sol (max):1 项分数修订。
相关对比
- DeepSeek V4 Pro 0424 vs GPT-6 Astra
- DeepSeek V4 Pro 0424 vs Gemini 4 Argon
- DeepSeek V4 Pro 0424 vs Claude Sonnet 5.5
- GPT-6.1 Sol vs Claude Fable 5.2
- GPT-6.1 Sol vs Gemini 4 Carbon
- GPT-6.1 Sol vs Claude Opus 5.5
- GPT-6.1 Sol vs GPT-6 Astra
- GPT-6.1 Sol vs Claude Fable 5.1
排行榜: 综合最佳大模型 · 最适合编程的大模型 · 最适合智能体的大模型 · 便宜好用的大模型
DeepSeek V4 Pro 0424 vs GPT-6.1 Sol:常见问题
- DeepSeek V4 Pro 0424 和 GPT-6.1 Sol 哪个更便宜?价格会变吗?
- 目前 DeepSeek V4 Pro 0424 更便宜:每百万 token $2.09 比 $4.00(按每 1 个输出 token 对应 3 个输入 token 混合计算)。两者均未公布价格调整。
- 智能体和工具调用用 DeepSeek V4 Pro 0424 还是 GPT-6.1 Sol 更好?
- GPT-6.1 Sol 在智能体任务方面得分更高(类别得分 64 比 60,50 为平均水平)。
- 处理长文档哪个更好?
- GPT-6.1 Sol 在长上下文基准测试上得分更高(66 比 60)。两者都约为 1.1M token(DeepSeek V4 Pro 0424 为 1,000,000,GPT-6.1 Sol 为 1,050,000)。
- 第三方基准测试对 DeepSeek V4 Pro 0424 和 GPT-6.1 Sol 的看法一致吗?
- 在两者都有结果的 5 项基准测试中,DeepSeek V4 Pro 0424 领先 0 项,GPT-6.1 Sol 领先 5 项。
- GPT-6.1 Sol 是什么时候发布的?
- GPT-6.1 Sol 于 2026年9月29日 发布。