信息情报站AI 情报 · 按工作日整理

变化追踪 · 历史快照

2026年11月15日 的追踪状态

当天快照,变化相对前一工作日。生成于 2026年11月16日 09:33。

历史快照

共 8 个每日快照,点任意一天查看当天的追踪状态。

当日快照 · 11月15日

相对前一工作日

API 价格

OpenRouter 公开报价 · 美元 / 百万 Token
模型输入输出
Google: Gemini 4 Hydrogen
google/gemini-4-hydrogen
$1.50±0$7.50±0
Google: Gemini 4 Hydrogen(批量)Google: Gemini 4 Hydrogen (batch)
google/gemini-4-hydrogen:batch
$0.75±0$3.75±0
Google: Gemini 4 Carbon
google/gemini-4-carbon
$2.00±0$10.00±0
Google: Gemini 4 Carbon(批量)Google: Gemini 4 Carbon (batch)
google/gemini-4-carbon:batch
$1.00±0$5.00±0
Anthropic: Claude Fable 5.2
anthropic/claude-fable-5.2
$8.00±0$40.00±0
Anthropic: Claude Fable 5.2(批处理)Anthropic: Claude Fable 5.2 (batch)
anthropic/claude-fable-5.2:batch
$4.00±0$20.00±0
Google: Gemini Omni 1.2 Flash
google/gemini-omni-1.2-flash
$1.50±0$9.00±0
Google: Gemini Omni 1.1 Flash
google/gemini-omni-1.1-flash
$1.50±0$9.00±0
Z.ai: GLM 5.3 FlashX Legal
z-ai/glm-5.3-flashx-legal
$0.37±0$1.25±0
OpenAI: GPT-6.1 Terra Pro
openai/gpt-6.1-terra-pro
$1.00±0$5.00±0
OpenAI:GPT-6.1 Terra Pro(批处理)OpenAI: GPT-6.1 Terra Pro (batch)
openai/gpt-6.1-terra-pro:batch
$0.50±0$2.50±0
OpenAI: GPT-6.1 Terra
openai/gpt-6.1-terra
$1.00±0$5.00±0
OpenAI: GPT-6.1 Terra(批处理)OpenAI: GPT-6.1 Terra (batch)
openai/gpt-6.1-terra:batch
$0.50±0$2.50±0
MoonshotAI: Kimi K3.1 Pro
moonshotai/kimi-k3.1-pro
$0.24-1.76$8.40-1.6
MoonshotAI: Kimi K3.1 Pro(批量)MoonshotAI: Kimi K3.1 Pro (batch)
moonshotai/kimi-k3.1-pro:batch
$0.70±0$6.00±0
DeepSeek: DeepSeek V4.1 Flash 1017
deepseek/deepseek-v4.1-flash-1017
$0.30+0.262$1.20+0.7
DeepSeek: DeepSeek V4.1 Flash 1017(批处理)DeepSeek: DeepSeek V4.1 Flash 1017 (batch)
deepseek/deepseek-v4.1-flash-1017:batch
$0.04-0.011$0.35+0.197
Google: Gemini 4 Argon
google/gemini-4-argon
$2.00±0$10.00±0
Google: Gemini 4 Argon(批量)Google: Gemini 4 Argon (batch)
google/gemini-4-argon:batch
$1.00±0$5.00±0
Anthropic: Claude Haiku 5.5
anthropic/claude-haiku-5.5
$0.10±0$0.50±0
Anthropic: Claude Haiku 5.5(批处理)Anthropic: Claude Haiku 5.5 (batch)
anthropic/claude-haiku-5.5:batch
$0.05±0$0.25±0
Google: Nano Banana 2.1
google/gemini-nano-banana-2.1
$1.50±0$7.50±0
Mistral: Mistral Large 4
mistralai/mistral-large-4-0
$0.70±0$2.10±0
OpenAI: GPT-6.1 Sol Pro
openai/gpt-6.1-sol-pro
$2.00±0$10.00±0
OpenAI:GPT-6.1 Sol Pro(批处理)OpenAI: GPT-6.1 Sol Pro (batch)
openai/gpt-6.1-sol-pro:batch
$1.00±0$5.00±0
OpenAI: GPT-6.1 Sol
openai/gpt-6.1-sol
$2.00±0$10.00±0
OpenAI: GPT-6.1 Sol(批处理)OpenAI: GPT-6.1 Sol (batch)
openai/gpt-6.1-sol:batch
$1.00±0$5.00±0
Anthropic: Claude Sonnet 5.5
anthropic/claude-sonnet-5.5
$2.00±0$10.00±0
Anthropic: Claude Sonnet 5.5(批处理)Anthropic: Claude Sonnet 5.5 (batch)
anthropic/claude-sonnet-5.5:batch
$1.00±0$5.00±0
Z.ai: GLM 5.3 Prime
z-ai/glm-5.3-prime
$2.80±0$8.80±0
Qwen: Qwen3.8 Max Prime
qwen/qwen3.8-max-prime
$4.00±0$12.00±0
OpenAI: GPT-6 Luna Pro
openai/gpt-6-luna-pro
$0.10±0$0.50±0
OpenAI:GPT-6 Luna Pro(批处理)OpenAI: GPT-6 Luna Pro (batch)
openai/gpt-6-luna-pro:batch
$0.05±0$0.25±0
OpenAI: GPT-6 Luna
openai/gpt-6-luna
$0.10±0$0.50±0
OpenAI: GPT-6 Luna(批处理)OpenAI: GPT-6 Luna (batch)
openai/gpt-6-luna:batch
$0.05±0$0.25±0
OpenAI: GPT-6 Sol Pro
openai/gpt-6-sol-pro
$2.00±0$10.00±0
OpenAI:GPT-6 Sol Pro(批处理)OpenAI: GPT-6 Sol Pro (batch)
openai/gpt-6-sol-pro:batch
$1.00±0$5.00±0
OpenAI: GPT-6 Sol
openai/gpt-6-sol
$2.00±0$10.00±0
OpenAI: GPT-6 Sol(批处理)OpenAI: GPT-6 Sol (batch)
openai/gpt-6-sol:batch
$1.00±0$5.00±0
Anthropic: Claude Opus 5.5
anthropic/claude-opus-5.5
$4.00±0$20.00±0

评测榜

排名变化相对上一工作日

LMArena文本(风格控制,综合)LMArena Text (style control, overall)

来源 ↗

指标:elo

  1. 1
    Gemini 4 Carbon (High) (+0)
    1526.182
  2. 2
    Gemini 4 Argon (High) (+0)
    1521.6933
  3. 3
    Muse Wonder (Max) (+0)
    1510.8561
  4. 4
    Claude Opus 5.5(高)Claude Opus 5.5 (High)(+0)
    1506.7849
  5. 5
    Claude Fable 5(高)Claude Fable 5 (High)(+0)
    1503.9797
  6. 6
    Claude Opus 4.6(高)Claude Opus 4.6 (High)(+0)
    1502.8295
  7. 7
    Claude Fable 5.1(最高)Claude Fable 5.1 (Max)(+0)
    1502.323
  8. 8
    Claude Opus 4.7(高)Claude Opus 4.7 (High)(+0)
    1501.1331
  9. 9
    Gemini 3.8 Flash (High) (+0)
    1496.6647
  10. 10
    Muse Spark 1.3 (Max) (+0)
    1496.5016

官方 Hugging Face 排行榜数据集,最新数据分割,综合类别。Official Hugging Face leaderboard-dataset, latest split, overall category.

Artificial Analysis智能指数Artificial Analysis Intelligence Index

来源 ↗

指标:intelligence_index

  1. 1
    Claude Fable 5.2(最高)Claude Fable 5.2 (Max)(+0)
    64.5636
  2. 2
    Claude Fable 5.2(超高)Claude Fable 5.2 (Xhigh)(+0)
    62.2167
  3. 3
    Claude Fable 5.2(高)Claude Fable 5.2 (High)(+0)
    61.8219
  4. 4
    Gemini 4 Carbon (High) (+0)
    60.0257
  5. 5
    Claude Opus 5.5(最高)Claude Opus 5.5 (Max)(+0)
    59.106
  6. 6
    Gemini 4 Carbon (Medium) (+0)
    58.2588
  7. 7
    Claude Opus 5.5 (Xhigh) (+0)
    57.4447
  8. 8
    GPT-6 Astra (Max) (+0)
    56.8539
  9. 9
    Claude Fable 5.2(中)Claude Fable 5.2 (Medium)(+0)
    56.4083
  10. 10
    GPT-6 Astra (Xhigh) (+0)
    55.3057

从官方公开排行榜页面数据中解析(无需 API 密钥)。官方 Data API 仍需密钥。Parsed from the official public leaderboard page payload (no API key). Official Data API remains key-gated.

LiveBench

来源 ↗

指标:category_mean

  1. 1
    gpt-5.5-xhigh (+0)
    80.5343
  2. 2
    claude-fable-5-xhigh-effort (+0)
    80.1342
  3. 3
    claude-opus-4-8-xhigh-effort (+0)
    79.8089
  4. 4
    gpt-5.5-high (+0)
    79.5625
  5. 5
    gpt-5.4-xhigh (+0)
    78.7983
  6. 6
    gemini-3.1-pro-preview-high (+0)
    78.151
  7. 7
    claude-opus-4-7-xhigh-effort (+0)
    77.0353
  8. 8
    claude-sonnet-5-xhigh-effort (+0)
    75.5064
  9. 9
    gemini-3.5-flash-high (+0)
    75.4373
  10. 10
    gpt-5.2-2025-12-11-high (+0)
    75.1918

官方 GitHub Pages 表格 table_2026_06_25.csv。由于 CSV 没有单独的官方总分列,总分取已公布分类列的平均值。Official GitHub Pages table table_2026_06_25.csv. Overall score is the mean of the published category columns, because the CSV has no single official overall column.

SWE-bench Verified

来源 ↗

指标:resolved_pct

  1. 1
    Sonar 基础智能体 + Claude 4.5 OpusSonar Foundation Agent + Claude 4.5 Opus(+0)
    79.2
  2. 2
    live-SWE-agent + Claude 4.5 Opus 中等(20251101)live-SWE-agent + Claude 4.5 Opus medium (20251101)(+0)
    79.2
  3. 3
    TRAE + Doubao-Seed-Code (+0)
    78.8
  4. 4
    live-SWE-agent + Gemini 3 Pro Preview (2025-11-18) (+0)
    77.4
  5. 5
    EPAM AI/Run Developer Agent v20250719 + Claude 4 Sonnet (+0)
    76.8
  6. 6
    Atlassian Rovo Dev (2025-09-02) (+0)
    76.8
  7. 7
    Claude 4.5 Opus(高)Claude 4.5 Opus (high)(+0)
    76.8
  8. 8
    ACoder (+0)
    76.4
  9. 9
    Gemini 3 Flash (high) (+0)
    75.8
  10. 10
    MiniMax M2.5(高)MiniMax M2.5 (high)(+0)
    75.8

官方 leaderboards.json,Verified 榜单,按解决百分比排名。Official leaderboards.json, Verified board, ranked by resolved percentage.

Aider多语言编程排行榜Aider polyglot coding leaderboard

来源 ↗

指标:percent_correct

  1. 1
    gpt-5 (high) (+0)
    88.0
  2. 2
    gpt-5 (medium) (+0)
    86.7
  3. 3
    o3-pro (high) (+0)
    84.9
  4. 4
    gemini-2.5-pro-preview-06-05(32k 思考)gemini-2.5-pro-preview-06-05 (32k think)(+0)
    83.1
  5. 5
    gpt-5 (low) (+0)
    81.3
  6. 6
    o3 (high) (+0)
    81.3
  7. 7
    grok-4 (high) (+0)
    79.6
  8. 8
    gemini-2.5-pro-preview-06-05(默认思考)gemini-2.5-pro-preview-06-05 (default think)(+0)
    79.1
  9. 9
    o3 (high) + gpt-4.1 (+0)
    78.2
  10. 10
    o3 (+0)
    76.9

官方 Aider 多语言 HTML 表格,按公布顺序排名。Official Aider polyglot HTML table, ranked as published.

Hugging Face 热门模型

按下载量代理,非官网 Trending
  1. 1171.2M-1.53e+06
  2. 2
    argmaxinc/whisperkit-coreml automatic-speech-recognition
    71.5M+3.92e+06
  3. 358.4M-8.29e+05
  4. 4
    BAAI/bge-small-en-v1.5 feature-extraction
    44.8M-5.05e+05
  5. 5
    openai-community/gpt2 text-generation
    33.9M+2.96e+05
  6. 633.3M+3.1e+05
  7. 7
    autogluon/chronos-2 time-series-forecasting
    33.2M-4.79e+05
  8. 831.5M-7.68e+05
  9. 9
    BAAI/bge-m3 sentence-similarity
    30.2M+2.84e+05
  10. 1025.8M-3.42e+05
  11. 1120.0M-1.87e+05
  12. 12
    google-t5/t5-small translation
    19.5M-1.81e+05
  13. 13
    Qwen/Qwen3-0.6B text-generation
    17.8M-7.23e+05
  14. 14
    openai/clip-vit-base-patch32 zero-shot-image-classification
    17.7M-4.15e+04
  15. 15
    microsoft/Florence-2-base image-text-to-text
    16.6M+8.34e+05
  16. 1616.2M-2.82e+04
  17. 17
    google/vit-base-patch16-224 image-classification
    15.7M-8.69e+04
  18. 18
    BAAI/bge-reranker-v2-m3 text-classification
    14.7M-406
  19. 19
    Qwen/Qwen3-14B text-generation
    13.0M+2.37e+04
  20. 20
    hexgrad/Kokoro-82M text-to-speech
    12.7M+5.62e+04

GitHub Trending 星标

当日 Trending 快照
  1. 125.9k+1.37e+03
  2. 29316
  3. 36477
  4. 45582+758
  5. 51525+611
  6. 652.1k+674
  7. 77206+1.14e+03
  8. 857.0k+656
  9. 917.9k+524
  10. 1015.9k
  11. 1111.9k+725
  12. 121707

GitHub 星标速度

官方 repos API · 日 / 七日变化

Star 数来自官方 GitHub 仓库 API。增长速度是与本系列先前快照的变化,而非第三方历史数据产品。只有一个数据点的日期暂无增长速度。Star counts come from the official GitHub repos API. Velocity is the change versus earlier snapshots in this series, not a third-party history product. A day with only one point has no velocity yet.

数据口径说明
API 价格
输入 / 输出价格来自 OpenRouter 公开模型 API(按其报价的每 token 美元价,页面折算为每百万 token)。
Hugging Face
Hugging Face API 不支持按 trending 排序,快照以下载量作为公开代理指标,并非官网 Trending 列表。
GitHub 星标
Trending 星标数来自每日 HTML 快照;github_star_series 另外每天记录官方仓库 API 的星标数,逐步形成增速曲线。
评测榜
Arena 取自官方 HF leaderboard 数据集;Artificial Analysis 取自公开模型页;LiveBench 取官方 CSV(按类别均值);SWE-bench Verified 取官方 JSON;Aider polyglot 取官方 HTML 表格。

设置

主题
涨跌颜色
默认语言(抓取的新闻与资料)

页头的国旗 / 地球图标可随时临时切换。设置只保存在本机浏览器(localStorage)。