anthropic 的网站图标

Anthropic: Claude Opus 4.6 (batch)

anthropic/claude-opus-4.6:batch

Opus 4.6 is Anthropic’s strongest model for coding and long-running professional tasks. It is built for agents that operate across entire workflows rather than single prompts, making it especially effective for large codebases, complex refactors, and multi-step debugging that unfolds over time. The model shows deeper contextual understanding, stronger problem decomposition, and greater reliability on hard engineering tasks than prior generations.

Beyond coding, Opus 4.6 excels at sustained knowledge work. It produces near-production-ready documents, plans, and analyses in a single pass, and maintains coherence across very long outputs and extended sessions. This makes it a strong default for tasks that require persistence, judgment, and follow-through, such as technical design, migration planning, and end-to-end project execution.

For users upgrading from earlier Opus versions, see our official migration guide here

模态
输入/输出价格
US$2.50 / US$12.50每 1M
上下文
1.0M
发布时间
2026年2月4日

提供商

不同公司托管同一模型。OpenRouter 根据您选择的路由模式将请求路由到其中一家公司:Balanced(价格 + 速度)、Nitro(速度最快)或 Floor(价格最低)。

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。

加权平均值

输入
US$0/百万 token
输出
US$0/百万 token
价格历史
当前选择无实际价格数据

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)。

AutoExacto 基准测试
GPQA DiamondTAU-BenchAnthropic87.1%76.0%Azure87.9%75.1%Claude Platform on AWS87.2%75.2%auto-routing87.6%74.1%Google Vertex87.5%74.1%Google Vertex86.9%74.3%
+1 家其他提供商

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。

正常运行时间数据尚不足,暂时无法显示。

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。

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: Anthropic: Claude Opus 4.6 (batch) (Artificial Analysis、Design Arena 和 Vals AI)
来源基准测试得分
Artificial AnalysisClaude Opus 4.6 (Non-reasoning, High) GPQA Diamond84.0%
Artificial AnalysisClaude Opus 4.6 (Non-reasoning, High) HLE19.1%
Artificial AnalysisClaude Opus 4.6 (Non-reasoning, High) IFBench44.6%
Artificial AnalysisClaude Opus 4.6 (Non-reasoning, High) τ²-Bench Telecom84.8%
Artificial AnalysisClaude Opus 4.6 (Non-reasoning, High) AA-LCR67.0%
Artificial AnalysisClaude Opus 4.6 (Non-reasoning, High) CritPt2.8%
Artificial AnalysisClaude Opus 4.6 (Non-reasoning, High) Terminal-Bench Hard48.5%
Artificial AnalysisClaude Opus 4.6 (Max) GPQA Diamond89.6%
Artificial AnalysisClaude Opus 4.6 (Max) HLE39.9%
Artificial AnalysisClaude Opus 4.6 (Max) IFBench53.1%
Artificial AnalysisClaude Opus 4.6 (Max) τ²-Bench Telecom92.1%
Artificial AnalysisClaude Opus 4.6 (Max) AA-LCR78.0%
Artificial AnalysisClaude Opus 4.6 (Max) CritPt12.6%
Artificial AnalysisClaude Opus 4.6 (Max) Terminal-Bench Hard46.2%
Design ArenaClaude Opus 4.6 Agents Arena Agenticgamedev Elo1143
Design ArenaClaude Opus 4.6 Agents Arena Androidnative Elo1180
Design ArenaClaude Opus 4.6 Agents Arena Blender3d Elo1008
Design ArenaClaude Opus 4.6 Agents Arena Codetovideo Elo1127
Design ArenaClaude Opus 4.6 Agents Arena Full Stack Elo1194
Design ArenaClaude Opus 4.6 Agents Arena Mobile Apps Elo1198
Design ArenaClaude Opus 4.6 Agents Arena Pdf Elo1191
Design ArenaClaude Opus 4.6 Agents Arena Spreadsheet Elo995
Design ArenaClaude Opus 4.6 Agents Arena Webapps Elo1186
Design ArenaClaude Opus 4.6 Models Arena 3D Elo1254
Design ArenaClaude Opus 4.6 Models Arena Asciiart Elo1236
Design ArenaClaude Opus 4.6 Models Arena Code Categories Elo1277
Design ArenaClaude Opus 4.6 Models Arena Data Visualization Elo1276
Design ArenaClaude Opus 4.6 Models Arena Game Development Elo1261
Design ArenaClaude Opus 4.6 Models Arena SVG Elo1204
Design ArenaClaude Opus 4.6 Models Arena UI Component Elo1272
Design ArenaClaude Opus 4.6 Models Arena Website Elo1282
Vals AIClaude Opus 4.6 (Thinking) AIME95.6%
Vals AIClaude Opus 4.6 (Thinking) CaseLaw v262.1%
Vals AIClaude Opus 4.6 (Thinking) CorpFin v267.0%
Vals AIClaude Opus 4.6 (Thinking) Finance Agent v1.160.0%
Vals AIClaude Opus 4.6 (Thinking) GPQA Diamond89.6%
Vals AIClaude Opus 4.6 (Thinking) LiveCodeBench84.7%
Vals AIClaude Opus 4.6 (Thinking) LegalBench85.3%
Vals AIClaude Opus 4.6 (Thinking) MedCode49.1%
Vals AIClaude Opus 4.6 (Thinking) MedQA95.4%
Vals AIClaude Opus 4.6 (Thinking) MedScribe86.1%
Vals AIClaude Opus 4.6 (Thinking) MMLU Pro89.1%
Vals AIClaude Opus 4.6 (Thinking) MMMU Pro83.9%
Vals AIClaude Opus 4.6 (Thinking) MortgageTax68.5%
Vals AIClaude Opus 4.6 (Thinking) ProofBench v150.0%
Vals AIClaude Opus 4.6 (Thinking) SAGE v151.6%
Vals AIClaude Opus 4.6 (Thinking) SWE-bench78.2%
Vals AIClaude Opus 4.6 (Thinking) TaxEval v276.0%
Vals AIClaude Opus 4.6 (Thinking) Terminal-Bench 2.058.4%
Vals AIClaude Opus 4.6 (Thinking) Vals Index66.0%
Vals AIClaude Opus 4.6 (Thinking) Vals Multimodal Index65.2%
Vals AIClaude Opus 4.6 (Thinking) Vibe Code Bench v1.153.5%

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。

使用此模型的应用尚无足够数据。

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

文档

在 OpenRouter 上使用 Anthropic: Claude Opus 4.6 (batch) 构建应用的指南、实用教程、博客文章和工具。

常见问题