1. 图表数据
  2. o3-mini-high tied for #1 across: Coding, Math & Hard Prompts
Arena表格图 · 180 个数据点 · Arena

o3-mini-high tied for #1 across: Coding, Math & Hard Prompts

Arena 发布的一张图背后的数字,用原生图表重画。对照原图,研究数据,再做你自己的。

从图片估算 80%全部图表数据
图表来自 Arena. 来源: x.com

简化重绘

由我们用下面的数字重新绘制。悬停、点按或用方向键查看。

这是简化重绘:它用我们自己的样式展示数字,不是原图的样子。强调色标出这张图的主角。表格里带 ~ 的是估算值。

数字

180 个点中前 10 个免费
分组系列数值来源置信
early-grok-3Overall1估算0.97
early-grok-3Overall w/ Style Control1估算0.97
early-grok-3Hard Prompts1估算0.97
early-grok-3Hard Prompts w/ Style Control1估算0.97
early-grok-3Coding~2估算0.30
early-grok-3Math1估算0.97
early-grok-3Creative Writing1估算0.97
early-grok-3Instruction Following1估算0.97
early-grok-3Longer Query1估算0.97
early-grok-3Multi-Turn1估算0.97

Max 数据,目前留邮箱免费看

查看全部 180 个数据点并下载文件

这张图的完整表格、JSON、CSV 和 agent skill 文件。它们属于 Max;原型阶段,留邮箱即可解锁。

解锁全部 35 家巨头。每周一封,随时退订。 看看能得到什么 · 登录

gemini-2.0-flash-thinking-exp-01-21Overall2估算0.97
gemini-2.0-flash-thinking-exp-01-21Overall w/ Style Control3估算0.97
gemini-2.0-flash-thinking-exp-01-21Hard Prompts1估算0.97
gemini-2.0-flash-thinking-exp-01-21Hard Prompts w/ Style Control1估算0.97
gemini-2.0-flash-thinking-exp-01-21Coding2估算0.97
gemini-2.0-flash-thinking-exp-01-21Math1估算0.97
gemini-2.0-flash-thinking-exp-01-21Creative Writing2估算0.97
gemini-2.0-flash-thinking-exp-01-21Instruction Following~2估算0.30
gemini-2.0-flash-thinking-exp-01-21Longer Query2估算0.97
gemini-2.0-flash-thinking-exp-01-21Multi-Turn1估算0.97
gemini-2.0-pro-exp-02-05Overall2估算0.97
gemini-2.0-pro-exp-02-05Overall w/ Style Control3估算0.97
gemini-2.0-pro-exp-02-05Hard Prompts1估算0.97
gemini-2.0-pro-exp-02-05Hard Prompts w/ Style Control1估算0.97
gemini-2.0-pro-exp-02-05Coding2估算0.97
gemini-2.0-pro-exp-02-05Math2估算0.97
gemini-2.0-pro-exp-02-05Creative Writing2估算0.97
gemini-2.0-pro-exp-02-05Instruction Following2估算0.97
gemini-2.0-pro-exp-02-05Longer Query~2估算0.30
gemini-2.0-pro-exp-02-05Multi-Turn1估算0.97
chatgpt-4o-latest-20250129Overall~5估算0.30
chatgpt-4o-latest-20250129Overall w/ Style Control~2估算0.30
chatgpt-4o-latest-20250129Hard Prompts4估算0.97
chatgpt-4o-latest-20250129Hard Prompts w/ Style Control2估算0.97
chatgpt-4o-latest-20250129Coding2估算0.97
chatgpt-4o-latest-20250129Math11估算0.97
chatgpt-4o-latest-20250129Creative Writing1估算0.97
chatgpt-4o-latest-20250129Instruction Following2估算0.97
chatgpt-4o-latest-20250129Longer Query~5估算0.30
chatgpt-4o-latest-20250129Multi-Turn1估算0.97
deepseek-r1Overall5估算0.97
deepseek-r1Overall w/ Style Control~8估算0.30
deepseek-r1Hard Prompts~4估算0.30
deepseek-r1Hard Prompts w/ Style Control2估算0.97
deepseek-r1Coding2估算0.97
deepseek-r1Math~5估算0.30
deepseek-r1Creative Writing~5估算0.30
deepseek-r1Instruction Following~5估算0.30
deepseek-r1Longer Query5估算0.97
deepseek-r1Multi-Turn1估算0.97
gemini-2.0-flash-001Overall5估算0.97
gemini-2.0-flash-001Overall w/ Style Control8估算0.97
gemini-2.0-flash-001Hard Prompts4估算0.97
gemini-2.0-flash-001Hard Prompts w/ Style Control9估算0.97
gemini-2.0-flash-001Coding2估算0.97
gemini-2.0-flash-001Math~5估算0.30
gemini-2.0-flash-001Creative Writing5估算0.97
gemini-2.0-flash-001Instruction Following5估算0.97
gemini-2.0-flash-001Longer Query5估算0.97
gemini-2.0-flash-001Multi-Turn3估算0.97
o1-2024-12-17Overall5估算0.97
o1-2024-12-17Overall w/ Style Control3估算0.97
o1-2024-12-17Hard Prompts2估算0.97
o1-2024-12-17Hard Prompts w/ Style Control~1估算0.30
o1-2024-12-17Coding2估算0.97
o1-2024-12-17Math1估算0.97
o1-2024-12-17Creative Writing~5估算0.30
o1-2024-12-17Instruction Following2估算0.97
o1-2024-12-17Longer Query2估算0.97
o1-2024-12-17Multi-Turn5估算0.97
o1-previewOverall8估算0.97
o1-previewOverall w/ Style Control7估算0.97
o1-previewHard Prompts5估算0.97
o1-previewHard Prompts w/ Style Control~4估算0.30
o1-previewCoding2估算0.97
o1-previewMath~8估算0.30
o1-previewCreative Writing~7估算0.30
o1-previewInstruction Following~6估算0.30
o1-previewLonger Query~5估算0.30
o1-previewMulti-Turn5估算0.97
o3-mini-highOverall8估算0.97
o3-mini-highOverall w/ Style Control7估算0.97
o3-mini-highHard Prompts5估算0.97
o3-mini-highHard Prompts w/ Style Control6估算0.97
o3-mini-highCoding1估算0.97
o3-mini-highMath1估算0.97
o3-mini-highCreative Writing11估算0.97
o3-mini-highInstruction Following5估算0.97
o3-mini-highLonger Query6估算0.97
o3-mini-highMulti-Turn5估算0.97
deepseek-v3Overall11估算0.97
deepseek-v3Overall w/ Style Control~14估算0.30
deepseek-v3Hard Prompts13估算0.97
deepseek-v3Hard Prompts w/ Style Control~14估算0.30
deepseek-v3Coding12估算0.97
deepseek-v3Math13估算0.97
deepseek-v3Creative Writing7估算0.97
deepseek-v3Instruction Following11估算0.97
deepseek-v3Longer Query6估算0.97
deepseek-v3Multi-Turn7估算0.97
qwen-plus-0125Overall11估算0.97
qwen-plus-0125Overall w/ Style Control13估算0.97
qwen-plus-0125Hard Prompts13估算0.97
qwen-plus-0125Hard Prompts w/ Style Control12估算0.97
qwen-plus-0125Coding12估算0.97
qwen-plus-0125Math~11估算0.30
qwen-plus-0125Creative Writing9估算0.97
qwen-plus-0125Instruction Following11估算0.97
qwen-plus-0125Longer Query8估算0.97
qwen-plus-0125Multi-Turn10估算0.97
gla-4.5-plus-0111Overall~11估算0.50
gla-4.5-plus-0111Overall w/ Style Control~14估算0.50
gla-4.5-plus-0111Hard Prompts~14估算0.50
gla-4.5-plus-0111Hard Prompts w/ Style Control~16估算0.50
gla-4.5-plus-0111Coding~16估算0.50
gla-4.5-plus-0111Math~15估算0.50
gla-4.5-plus-0111Creative Writing~8估算0.50
gla-4.5-plus-0111Instruction Following~13估算0.50
gla-4.5-plus-0111Longer Query~11估算0.50
gla-4.5-plus-0111Multi-Turn~9估算0.50
gemini-2.0-flash-lite-preview-02-05Overall11估算0.97
gemini-2.0-flash-lite-preview-02-05Overall w/ Style Control13估算0.97
gemini-2.0-flash-lite-preview-02-05Hard Prompts12估算0.97
gemini-2.0-flash-lite-preview-02-05Hard Prompts w/ Style Control13估算0.97
gemini-2.0-flash-lite-preview-02-05Coding~12估算0.30
gemini-2.0-flash-lite-preview-02-05Math13估算0.97
gemini-2.0-flash-lite-preview-02-05Creative Writing~5估算0.30
gemini-2.0-flash-lite-preview-02-05Instruction Following12估算0.97
gemini-2.0-flash-lite-preview-02-05Longer Query9估算0.97
gemini-2.0-flash-lite-preview-02-05Multi-Turn13估算0.97
o3-miniOverall12估算0.97
o3-miniOverall w/ Style Control12估算0.97
o3-miniHard Prompts4估算0.97
o3-miniHard Prompts w/ Style Control~4估算0.30
o3-miniCoding2估算0.97
o3-miniMath~1估算0.30
o3-miniCreative Writing20估算0.97
o3-miniInstruction Following10估算0.97
o3-miniLonger Query~6估算0.30
o3-miniMulti-Turn11估算0.97
step-2-16k-exp-202412Overall12估算0.97
step-2-16k-exp-202412Overall w/ Style Control17估算0.97
step-2-16k-exp-202412Hard Prompts14估算0.97
step-2-16k-exp-202412Hard Prompts w/ Style Control14估算0.97
step-2-16k-exp-202412Coding14估算0.97
step-2-16k-exp-202412Math~12估算0.30
step-2-16k-exp-202412Creative Writing5估算0.97
step-2-16k-exp-202412Instruction Following~14估算0.30
step-2-16k-exp-202412Longer Query10估算0.97
step-2-16k-exp-202412Multi-Turn13估算0.97
o1-miniOverall12估算0.97
o1-miniOverall w/ Style Control17估算0.97
o1-miniHard Prompts11估算0.97
o1-miniHard Prompts w/ Style Control11估算0.97
o1-miniCoding3估算0.97
o1-miniMath6估算0.97
o1-miniCreative Writing29估算0.97
o1-miniInstruction Following12估算0.97
o1-miniLonger Query10估算0.97
o1-miniMulti-Turn11估算0.97
gemini-1.5-pro-002Overall12估算0.97
gemini-1.5-pro-002Overall w/ Style Control13估算0.97
gemini-1.5-pro-002Hard Prompts14估算0.97
gemini-1.5-pro-002Hard Prompts w/ Style Control14估算0.97
gemini-1.5-pro-002Coding16估算0.97
gemini-1.5-pro-002Math~13估算0.30
gemini-1.5-pro-002Creative Writing~7估算0.30
gemini-1.5-pro-002Instruction Following13估算0.97
gemini-1.5-pro-002Longer Query13估算0.97
gemini-1.5-pro-002Multi-Turn15估算0.97
glm-4-plus-0111Overall~11估算0.50
glm-4-plus-0111Overall w/ Style Control~13估算0.50
glm-4-plus-0111Hard Prompts~14估算0.50
glm-4-plus-0111Hard Prompts w/ Style Control~16估算0.50
glm-4-plus-0111Coding~16估算0.50
glm-4-plus-0111Math~15估算0.50
glm-4-plus-0111Creative Writing~8估算0.50
glm-4-plus-0111Instruction Following~8估算0.50
glm-4-plus-0111Longer Query~11估算0.50
glm-4-plus-0111Multi-Turn~10估算0.50
下载这张图的数据

Max 订阅者可以通过 API 取同样的数据。 API 文档

API · Max

用 API 调用这张图

Max 套餐可以取这张图的数字(每个数字都带来源),格式为 JSON 或 CSV,附重画规格;也可以让智能体通过 MCP 调用。图表 id:b9fdb5e6af79。

路径返回套餐
GET/api/v1/charts/b9fdb5e6af79标题、发布者、日期、图型、数据点数;用 Max 密钥时附数据链接开放
GET/api/v1/charts/b9fdb5e6af79/data每个数据点及其来源和置信度,加上重画规格。加 ?format=csv 得到 CSVMax
GET/api/v1/charts/b9fdb5e6af79/render?view=1本页使用的只读绘图视图开放
POST/mcp工具 get_chart_data 加这个 id,供 Claude、Cursor 等智能体使用Max
curlMax
curl -s https://viral-chart.com/api/v1/charts/b9fdb5e6af79/data \
  -H "Authorization: Bearer $VIRALCHART_KEY"
Claude CodeMax
claude mcp add --transport http viralchart https://viral-chart.com/mcp \
  --header "Authorization: Bearer $VIRALCHART_KEY"

每个密钥每分钟 60 次,超出返回 429 和 Retry-After。每个账号最多 5 个密钥,在账户页创建,只显示一次。没有 Max 时数据接口返回 402,并附套餐链接。这些数字描述的是发布者拥有的图,每个文件都带授权说明。

做一张这样的图

这张图的形式,你的数字。

实时预览。所有内容都留在你的设备上。

选择形式

简化形式:用我们自己的样式,不是原图的样子。你的数字从这张图的数据开始。

Arena 的更多数据

全部图表数据

常见问题

这些数字有多准?

它们由两遍独立读图估算得出,置信度 0.80(0 到 1)。图上印着的数字读得准,从坐标轴读出的值可能有几个百分点的偏差。

这些数据可以直接用吗?

这些数字描述的是 Arena 发布的图表。图片、版式和样式归他们所有。使用数字时请注明发布者,图片请用你自己的数据来做。

怎么拿到文件?

JSON、CSV 和 agent skill 文件在本页的邮箱解锁区。原型里它们是静态文件:/api/charts/b9fdb5e6af79.json 和 /api/charts/b9fdb5e6af79.skill.md。