跳转到内容
Alpha
  • 模型
  • 排行榜
  • Ori
联系销售

页脚

OpenRouterOpenRouter
© 2026 OpenRouter, Inc

产品

  • 聊天
  • 排行榜
  • 基准测试
  • 应用
  • 发现
  • 模型
  • Ori
  • 合集
  • 提供商
  • 工具
  • 定价
  • Business
  • Enterprise
  • 实验室

公司

  • 关于
  • 博客
  • 客户
  • 招聘
    招聘
  • 隐私
  • 服务条款
  • 信任中心
  • 支持
  • 兼容 OR
  • 数据
  • 品牌

开发者

  • 文档
  • API 参考
  • 开发者平台
  • 状态
  • AI 站点地图

连接

  • Discord
  • GitHub
  • LinkedIn
  • X
  • YouTube

The free Inkling endpoint is only available for use with agentic harnesses. Do not upload any confidential information or personal data (e.g., voices and images of people's faces). Your usage of this free endpoint, including prompts and outputs, is logged and used to improve Thinking Machines Lab's models, products, and services. The logged session data will be disassociated from your account and other persistent identifiers before being used for these purposes.

By using this free endpoint, you agree to the TML Free Research API Terms of ServiceOpens in new tab. For more information about Thinking Machines Lab's data processing practices, see this Privacy NoticeOpens in new tab.

thinkingmachines 的网站图标

Thinking Machines: Inkling

thinkingmachines/inkling

模型权重
比较

Inkling is an open-weight multimodal mixture-of-experts model from Thinking Machines Lab, with 41B active parameters out of 975B total. It is designed for general-purpose reasoning, coding, agentic and tool-use systems, retrieval-augmented generation, instruction following, and multilingual conversational applications. Its native image and audio understanding supports multimodal analysis alongside text.

模态
输入 / 输出价格
US$1 / US$4.05每 1M
上下文
524K
发布时间
2026年7月17日
比较
提供商定价性能正常运行时间基准测试应用活动常见问题探索

提供商

此模型由一家提供商托管。OpenRouter 将每个请求直接转发给该提供商,无需做出路由决策。

定价

客户为此模型实际支付的平均价格,与提供商公布的价格并列展示。由于缓存和折扣,实际支付的价格往往远低于标价。

性能

吞吐量是模型生成内容的速度(每秒 token 数,越高越好)。延迟是完整的往返时间(越低越好)。TTFT 是首个 token 的响应时间,即您看到内容出现前需要等待的时间(越低越好)。

正常运行时间

正常运行时间是指过去 3 天中至少有一家提供商响应请求的时间百分比。可用性是指成功提供推理服务的时间百分比。OpenRouter 会持续监控,并在某家提供商返回错误时使用次优提供商。

基准测试

标准化评测得分,以及此模型针对共用基准测试提示词的实际输出。百分比越高越好,每项排名显示此模型在该来源所评测的模型中的位置。

基准测试得分摘要: Thinking Machines: Inkling (Artificial Analysis、Design Arena 和 Vals AI)
来源基准测试得分
Artificial AnalysisInkling (Xhigh) Intelligence Index12.5
Artificial AnalysisInkling (Xhigh) Coding Index52.1
Artificial AnalysisInkling (Xhigh) Agentic Index22.5
Artificial AnalysisInkling (Xhigh) GPQA Diamond87.2%
Artificial AnalysisInkling (Xhigh) HLE31.9%
Artificial AnalysisInkling (Xhigh) AA-LCR77.3%
Artificial AnalysisInkling (Xhigh) τ-Bench Banking29.1%
Artificial AnalysisInkling (Xhigh) GDPval-AA29.0%
Artificial AnalysisInkling (Xhigh) CritPt5.4%
Artificial AnalysisInkling (Xhigh) SciCode47.0%
Artificial AnalysisInkling (Xhigh) Terminal-Bench 4.01.0%
Artificial AnalysisInkling (Xhigh) Terminal-Bench 2.155.1%
Artificial AnalysisInkling (Xhigh) AA-Omniscience Accuracy39.3%
Artificial AnalysisInkling (Xhigh) AA-Omniscience Non-Hallucination Rate32.8%
Design ArenaInkling Agents Arena Webapps Elo1180
Design ArenaInkling Models Arena 3D Elo1157
Design ArenaInkling Models Arena Asciiart Elo1092
Design ArenaInkling Models Arena Code Categories Elo1204
Design ArenaInkling Models Arena Data Visualization Elo1157
Design ArenaInkling Models Arena Game Development Elo1165
Design ArenaInkling Models Arena SVG Elo1106
Design ArenaInkling Models Arena UI Component Elo1196
Design ArenaInkling Models Arena Website Elo1223
Vals AIInkling Code Migration v1.18.8%
Vals AIInkling CorpFin v268.6%
Vals AIInkling CyberBench v1.166.8%
Vals AIInkling EMB v1.141.0%
Vals AIInkling Finance Agent v2.149.9%
Vals AIInkling Vals Finance Index38.2%
Vals AIInkling GPQA Diamond87.1%
Vals AIInkling HLAB-Vals-1.13.8%
Vals AIInkling IOI14.9%
Vals AIInkling LiveCodeBench85.5%
Vals AIInkling LegalBench82.9%
Vals AIInkling Vals Legal Index54.2%
Vals AIInkling Legal Research Bench28.4%
Vals AIInkling MedCode41.2%
Vals AIInkling MedScribe85.4%
Vals AIInkling MMLU Pro86.3%
Vals AIInkling MMMU Pro78.5%
Vals AIInkling MortgageTax64.0%
Vals AIInkling ProgramBench0.0%
Vals AIInkling ProofBench v1.10.0%
Vals AIInkling Public Benefits Bench v1.158.5%
Vals AIInkling SAGE v136.6%
Vals AIInkling SkillsBench26.1%
Vals AIInkling SRE Bench0.0%
Vals AIInkling SWE-bench77.6%
Vals AIInkling Tax Agent Bench15.3%
Vals AIInkling TaxEval v275.3%
Vals AIInkling Terminal-Bench 4.00.5%
Vals AIInkling Vals Index29.0%
Vals AIInkling Vals Multimodal Index49.4%
Vals AIInkling Vibe Code Bench 1-1007.3%
Vals AIInkling Vibe Code Bench v1.119.2%

应用

向此模型发送最多流量的公开应用。这能很好地反映实际生产工作负载的情况,也能提示此模型最适合哪些用例。

活动

此模型的 token 用量和请求流量随时间的变化。

API

可直接用于调用此模型的代码。OpenRouter 的 API 与 OpenAI 兼容——大多数 SDK 只需替换基础 URL 即可使用。切换模型时,唯一需要更改的是下方的模型标识符。

目前无法通过 EU 区域内路由使用。 区域内路由文档

探索更多模型

具备视觉能力的 AI 模型:用于图像理解的多模态大语言模型集合AI 模型排行榜排名
延迟 / 吞吐量
US$1.00US$4.05US$0.170.48s63 tps
95.70%
吞吐量
63token/秒

P50,所有提供商中最佳

延迟
0.48秒

P50,最佳提供商

AutoExacto 基准测试
GPQA DiamondTAU-BenchDeepInfra85.1%68.0%
正常运行时间(3 天)模型可访问。请求已路由至提供商。
100.00%
可用性(3 天)模型通过任意提供商返回了推理结果。错误和空响应均计为失败。
94.01%

过去 3 天的可用性

最近 72 小时
可用性 94.01%
3 天前2 天前昨天现在

过去 24 小时的可用性

OpenRouter Availability
90.79%

当上游提供商发生错误时,如果您的请求筛选条件允许,我们可以通过路由到另一家健康的提供商来恢复服务。您可以通过以下接口以编程方式访问各提供商的正常运行时间数据: 端点 API。了解更多有关我们的 负载均衡和提供商路由选项。

1.
https://nousresearch.com 的网站图标
Hermes Agent
Hermes Agent is an open-source, self-improving AI agent by Nous Research that runs persistently with memory across sessions, and builds reusable skills from experience. It comes with 40+ built-in tools, including web search, browser automation, and vision, plus scheduled automations and subagents.
2.7Btoken
2.
https://openai.com/codex 的网站图标
Codex
A coding agent that helps you build and ship with AI
2.16Btoken
3.
https://docs.all-hands.dev/ 的网站图标
OpenHands
AI coding agent that can run commands, browse the web, call APIs
661Mtoken
4.
https://afterquery.com/ 的网站图标
ale-evol
Applied research lab for frontier models
377Mtoken
5.
https://llm.datasette.io/ 的网站图标
LLM
A CLI and Python library for LLMs
315Mtoken

常见问题

Inkling is an open-weight multimodal mixture-of-experts model from Thinking Machines Lab, with 41B active parameters out of 975B total. It is designed for general-purpose reasoning, coding, agentic and tool-use systems, retrieval-augmented generation, instruction following, and multilingual conversational applications.

Inkling 的费用为 US$1.00/百万输入 token 和 US$4.05/百万输出 token,Cache Read,价格为 US$0.17/M tokens 单独计费。

Inkling 的上下文窗口为 524,288 token。

是的。Inkling 接受 tools 和 tool_choice,以进行函数调用。 它不支持 response_format,因此不会强制输出 JSON 格式。

Inkling 接受 文本、图像和音频 作为输入,并返回 文本。

Inkling Small 是 Thinking Machines 推出的另一款 文本 模型。

Inkling 于 2026年7月17日 发布。

更多模型,来自 thinkingmachines

Inkling Small

Inkling Small is an open-weight multimodal mixture-of-experts model from Thinking Machines Lab, with 12B active parameters out of 276B total. It is positioned as the smaller, more efficient member of the Inkling family and is suited for reasoning, coding, agentic workflows, retrieval-augmented generation, instruction following, and multilingual conversation.

文本524K 上下文US$0.45 / US$1.20
Inkling Small (free)

Inkling Small is an open-weight multimodal mixture-of-experts model from Thinking Machines Lab, with 12B active parameters out of 276B total. It is positioned as the smaller, more efficient member of the Inkling family and is suited for reasoning, coding, agentic workflows, retrieval-augmented generation, instruction following, and multilingual conversation.

文本1.0M 上下文Free
Inkling (free)

Inkling is an open-weight multimodal mixture-of-experts model from Thinking Machines Lab, with 41B active parameters out of 975B total. It is designed for general-purpose reasoning, coding, agentic and tool-use systems, retrieval-augmented generation, instruction following, and multilingual conversational applications. Its native image and audio understanding supports multimodal analysis alongside text.

文本1.0M 上下文Free