{ "displayName": "Baidu Qianfan", "baseUrl": "https://qianfan.baidubce.com/v2", "apiKeyTemplate": "bce-v3/ALTAK-xxxxxxxxxxxxxxxxxxxxx/xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx", "tokenKeyTemplate": "bce-v3/ALTAK-xxxxxxxxxxxxxxxxxxxxxxxx/xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx", "models": [ { "id": "glm-5.3-token", "name": "GLM-5.3 (TokenPlan)", "tooltip": "百度千帆 Token Plan 个人专属接入点中的 GLM-5.3。智谱最新旗舰模型，编程与网络安全能力突出，可靠交付生产级代码。支持 1M 上下文窗口，长任务稳定可用。", "provider": "baidu-token", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/personal", "model": "glm-5.3", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "low"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "RMB": [8, 28, 2] } }, { "id": "glm-5.3-flash-token", "name": "GLM-5.3-Flash (TokenPlan)", "tooltip": "百度千帆 Token Plan 个人专属接入点中的 GLM-5.3-Flash。GLM-5 系列首个原生多模态模型，以极致低成本架构实现超越 GLM-5.2 的更强智能。在保持精准长上下文能力的同时，大幅降低了长上下文服务成本。", "provider": "baidu-token", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/personal", "model": "glm-5.3-flash", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "low"], "capabilities": { "toolCalling": true, "imageInput": true }, "tokenPricing": { "RMB": [0.8, 2.8, 0.23] } }, { "id": "glm-5.2-token", "name": "GLM-5.2 (TokenPlan)", "tooltip": "百度千帆 Token Plan 个人专属接入点中的 GLM-5.2。支持真正可用的 1M 上下文，在长程任务中继续保持领先。", "provider": "baidu-token", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/personal", "model": "glm-5.2", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "none"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "RMB": [8, 28, 2] } }, { "id": "glm-5.1-token", "name": "GLM-5.1 (TokenPlan)", "tooltip": "百度千帆 Token Plan 个人专属接入点中的 GLM-5.1。代码能力大大增强，长程任务显著提升，交付工程级成果。", "provider": "baidu-token", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/personal", "model": "glm-5.1", "maxInputTokens": 168000, "maxOutputTokens": 32000, "thinking": ["enabled", "disabled"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "pricing": { "RMB": [6, 24, 1.3] }, "tiers": [{ "contextSizeMin": 32001, "pricing": { "RMB": [8, 28, 2] } }] } }, { "id": "deepseek-v4-pro-token", "name": "DeepSeek-V4-Pro (TokenPlan)", "tooltip": "百度千帆 Token Plan 个人专属接入点中的 DeepSeek-V4-Pro。支持百万级超长上下文，在 Agent 能力、世界知识和推理性能上均实现国内与开源领域的领先。", "provider": "baidu-token", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/personal", "model": "deepseek-v4-pro", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "none"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "RMB": [12, 24, 1] } }, { "id": "deepseek-v4-pro-0813-token", "name": "DeepSeek-V4-Pro-0813 (TokenPlan)", "tooltip": "百度千帆 Token Plan 个人专属接入点中的 DeepSeek-V4-Pro-0813（1.8x 抵扣系数）。DeepSeek-V4-Pro 正式版本，大幅增强 Agent 能力，尤其在生产环境中表现突出。", "provider": "baidu-token", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/personal", "model": "deepseek-v4-pro-0813", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "low", "none"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "pricing": { "RMB": [4.5, 13.5, 0.45] }, "tiers": [{ "cron": "* 8-21 * * *", "pricing": { "RMB": [9, 27, 0.9] } }] } }, { "id": "deepseek-v4-flash-token", "name": "DeepSeek-V4-Flash (TokenPlan)", "tooltip": "百度千帆 Token Plan 个人专属接入点中的 DeepSeek-V4-Flash。高效轻量化模型，支持百万超长上下文能力。", "provider": "baidu-token", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/personal", "model": "deepseek-v4-flash", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "none"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "RMB": [1, 2, 0.2] } }, { "id": "deepseek-v4-flash-0731-token", "name": "DeepSeek-V4-Flash-0731 (TokenPlan)", "tooltip": "百度千帆 Token Plan 个人专属接入点中的 DeepSeek-V4-Flash-0731。高效轻量化模型，DeepSeek-V4-Flash 官方正式版本，取代预览版，大幅增强 Agent 能力。", "provider": "baidu-token", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/personal", "model": "deepseek-v4-flash-0731", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "low", "none"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "pricing": { "RMB": [1.5, 4.5, 0.15] }, "tiers": [{ "cron": "* 8-21 * * *", "pricing": { "RMB": [3, 9, 0.3] } }] } }, { "id": "kimi-k2.6-token", "name": "Kimi-K2.6 (TokenPlan)", "tooltip": "百度千帆 Token Plan 个人专属接入点中的 Kimi-K2.6。具备更强更稳的长程代码编写能力，同时支持文本、图片输入。", "provider": "baidu-token", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/personal", "model": "kimi-k2.6", "maxInputTokens": 224000, "maxOutputTokens": 32000, "thinking": ["enabled", "disabled"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "RMB": [6.5, 27, 1.1] } }, { "id": "deepseek-v4-pro-token-enterprise", "name": "DeepSeek-V4-Pro (TokenPlanEnterprise)", "tooltip": "百度千帆 Token Plan 企业版专属接入点中的 DeepSeek-V4-Pro。支持百万级超长上下文，在 Agent 能力、世界知识和推理性能上均实现国内与开源领域的领先。", "provider": "baidu-token-enterprise", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/team", "model": "deepseek-v4-pro", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "none"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "RMB": [12, 24, 1] } }, { "id": "deepseek-v4-pro-0813-token-enterprise", "name": "DeepSeek-V4-Pro-0813 (TokenPlanEnterprise)", "tooltip": "百度千帆 Token Plan 企业版专属接入点中的 DeepSeek-V4-Pro-0813（1.8x 抵扣系数）。DeepSeek-V4-Pro 正式版本，大幅增强 Agent 能力，尤其在生产环境中表现突出。", "provider": "baidu-token-enterprise", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/team", "model": "deepseek-v4-pro-0813", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "low", "none"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "pricing": { "RMB": [4.5, 13.5, 0.45] }, "tiers": [{ "cron": "* 8-21 * * *", "pricing": { "RMB": [9, 27, 0.9] } }] } }, { "id": "deepseek-v4-flash-token-enterprise", "name": "DeepSeek-V4-Flash (TokenPlanEnterprise)", "tooltip": "百度千帆 Token Plan 企业版专属接入点中的 DeepSeek-V4-Flash。高效轻量化模型，支持百万超长上下文能力。", "provider": "baidu-token-enterprise", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/team", "model": "deepseek-v4-flash", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "none"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "RMB": [1, 2, 0.2] } }, { "id": "deepseek-v4-flash-0731-token-enterprise", "name": "DeepSeek-V4-Flash-0731 (TokenPlanEnterprise)", "tooltip": "百度千帆 Token Plan 企业版专属接入点中的 DeepSeek-V4-Flash-0731。DeepSeek-V4-Flash 官方正式版本，取代预览版，大幅增强 Agent 能力。", "provider": "baidu-token-enterprise", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/team", "model": "deepseek-v4-flash-0731", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "low", "none"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "pricing": { "RMB": [1.5, 4.5, 0.15] }, "tiers": [{ "cron": "* 8-21 * * *", "pricing": { "RMB": [3, 9, 0.3] } }] } }, { "id": "glm-5.3-token-enterprise", "name": "GLM-5.3 (TokenPlanEnterprise)", "tooltip": "百度千帆 Token Plan 企业版专属接入点中的 GLM-5.3。智谱AI推出的新一代大语言模型，以极致后训练 Scaling 实现能力跃迁。", "provider": "baidu-token-enterprise", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/team", "model": "glm-5.3", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "low"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "RMB": [8, 28, 2] } }, { "id": "glm-5.3-flash-token-enterprise", "name": "GLM-5.3-Flash (TokenPlanEnterprise)", "tooltip": "百度千帆 Token Plan 企业版专属接入点中的 GLM-5.3-Flash。GLM-5 系列首个原生多模态模型，采用稀疏注意力与线性注意力混合架构，以极致低成本实现超越 GLM-5.2 的更强智能。", "provider": "baidu-token-enterprise", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/team", "model": "glm-5.3-flash", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "low"], "capabilities": { "toolCalling": true, "imageInput": true }, "tokenPricing": { "RMB": [0.8, 2.8, 0.23] } }, { "id": "glm-5.2-token-enterprise", "name": "GLM-5.2 (TokenPlanEnterprise)", "tooltip": "百度千帆 Token Plan 企业版专属接入点中的 GLM-5.2。支持真正可用的 1M 上下文，在长程任务中继续保持领先。", "provider": "baidu-token-enterprise", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/team", "model": "glm-5.2", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "none"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "RMB": [8, 28, 2] } }, { "id": "glm-5.1-token-enterprise", "name": "GLM-5.1 (TokenPlanEnterprise)", "tooltip": "百度千帆 Token Plan 企业版专属接入点中的 GLM-5.1。代码能力大大增强，长程任务显著提升，交付工程级成果。", "provider": "baidu-token-enterprise", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/team", "model": "glm-5.1", "maxInputTokens": 168000, "maxOutputTokens": 32000, "thinking": ["enabled", "disabled"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "pricing": { "RMB": [6, 24, 1.3] }, "tiers": [{ "contextSizeMin": 32001, "pricing": { "RMB": [8, 28, 2] } }] } }, { "id": "glm-5-token-enterprise", "name": "GLM-5 (TokenPlanEnterprise)", "tooltip": "百度千帆 Token Plan 企业版专属接入点中的 GLM-5。智谱新一代旗舰基座模型，面向 Agentic Engineering 打造，能够在复杂系统工程与长程 Agent 任务中提供可靠生产力。", "provider": "baidu-token-enterprise", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/team", "model": "glm-5", "maxInputTokens": 168000, "maxOutputTokens": 32000, "thinking": ["enabled", "disabled"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "pricing": { "RMB": [4, 18, 1] }, "tiers": [{ "contextSizeMin": 32001, "pricing": { "RMB": [6, 22, 1.5] } }] } }, { "id": "kimi-k2.6-token-enterprise", "name": "Kimi-K2.6 (TokenPlanEnterprise)", "tooltip": "百度千帆 Token Plan 企业版专属接入点中的 Kimi-K2.6。具备更强更稳的长程代码编写能力，同时支持文本、图片与视频输入，思考与非思考模式，对话与 Agent 任务。", "provider": "baidu-token-enterprise", "sdkMode": "anthropic", "baseUrl": "https://qianfan.baidubce.com/anthropic/tokenplan/team", "model": "kimi-k2.6", "maxInputTokens": 224000, "maxOutputTokens": 32000, "thinking": ["enabled", "disabled"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "RMB": [6.5, 27, 1.1] } }, { "id": "deepseek-v4-flash", "name": "DeepSeek-V4-Flash (PayGo)", "tooltip": "DeepSeek-V4 拥有百万字超长上下文，在 Agent 能力、世界知识和推理性能上均实现国内与开源领域的领先。DeepSeek-V4-Flash 是 DeepSeek-V4 系列的轻量高效版本，在保持 1M 超长上下文能力的同时，通过更小的模型参数与激活规模，提供更为快捷、经济的 API 服务。", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "none"], "thinkingFormat": "object", "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "RMB": [1, 2, 0.2] } }, { "id": "deepseek-v4-flash-0731", "name": "DeepSeek-V4-Flash-0731 (PayGo)", "tooltip": "DeepSeek-V4-Flash 官方正式版本，取代预览版，大幅增强 Agent 能力。", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "low", "none"], "thinkingFormat": "object", "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "pricing": { "RMB": [0.5, 1, 0.1] }, "tiers": [{ "cron": "* 8-21 * * *", "pricing": { "RMB": [1, 2, 0.2] } }] } }, { "id": "deepseek-v4-pro", "name": "DeepSeek-V4-Pro (PayGo)", "tooltip": "DeepSeek-V4 拥有百万字超长上下文，在 Agent 能力、世界知识和推理性能上均实现国内与开源领域的领先。", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "none"], "thinkingFormat": "object", "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "RMB": [12, 24, 1] } }, { "id": "deepseek-v4-pro-0813", "name": "DeepSeek-V4-Pro-0813 (PayGo)", "tooltip": "DeepSeek-V4-Pro 正式版本，大幅增强 Agent 能力，尤其在生产环境中表现突出。", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "low", "none"], "thinkingFormat": "object", "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "pricing": { "RMB": [6, 12, 0.5] }, "tiers": [{ "cron": "* 8-21 * * *", "pricing": { "RMB": [12, 24, 1] } }] } }, { "id": "glm-5", "name": "GLM-5 (PayGo)", "tooltip": "GLM-5 是智谱新一代的旗舰基座模型，面向 Agentic Engineering 打造，能够在复杂系统工程与长程 Agent 任务中提供可靠生产力。在 Coding 与 Agent 能力上，GLM-5 取得开源 SOTA 表现，在真实编程场景的使用体感逼近 Claude Opus 4.5，擅长复杂系统工程与长程 Agent 任务，是通用 Agent 助手的理想基座。", "maxInputTokens": 168000, "maxOutputTokens": 32000, "thinking": ["enabled", "disabled"], "thinkingFormat": "object", "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "pricing": { "RMB": [4, 18, 1] }, "tiers": [{ "contextSizeMin": 32001, "pricing": { "RMB": [6, 22, 1.5] } }] } }, { "id": "glm-5.1", "name": "GLM-5.1 (PayGo)", "tooltip": "GLM-5.1 是智谱新一代旗舰基座，长程任务显著提升，可自主工作长达8小时，闭环交付工程级成果，整体表现对齐Claude Opus 4.6。", "maxInputTokens": 168000, "maxOutputTokens": 32000, "thinking": ["enabled", "disabled"], "thinkingFormat": "object", "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "pricing": { "RMB": [6, 24, 1.3] }, "tiers": [{ "contextSizeMin": 32001, "pricing": { "RMB": [8, 28, 2] } }] } }, { "id": "glm-5.2", "name": "GLM-5.2 (PayGo)", "tooltip": "GLM-5.2 是面向长任务时代的旗舰模型。一次任务即可完成“从需求到多端可部署产物”的完整开发链路。", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "none"], "thinkingFormat": "object-none", "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "RMB": [8, 28, 2] } }, { "id": "glm-5.3", "name": "GLM-5.3 (PayGo)", "tooltip": "GLM-5.3是智谱AI推出的新一代大语言模型，以极致后训练 Scaling 实现能力跃迁。", "contextSize": [1000000, 512000, 400000, 256000, 192000], "maxInputTokens": 936000, "maxOutputTokens": 64000, "reasoningEffort": ["high", "max", "low"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "RMB": [8, 28, 2] } }, { "id": "kimi-k2.6", "name": "Kimi-K2.6 (PayGo)", "tooltip": "Kimi K2.6 是 Kimi 最新最智能的模型，Kimi K2.6 的通用 Agent、代码、视觉理解等综合能力得到全面提升，其中在博士级难度的完整版人类最后的考试（Humanity’s Last Exam）、在考察模型真实软件工程能力的 SWE-Bench Pro、评估 Agent 深度检索能力的 DeepSearchQA 等基准测试中均取得行业领先的成绩，同时支持文本、图片与视频输入，思考与非思考模式，对话与 Agent 任务。", "maxInputTokens": 224000, "maxOutputTokens": 32000, "thinking": ["enabled", "disabled"], "thinkingFormat": "object", "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "RMB": [6.5, 27, 1.1] } }, { "id": "ernie-5.0", "name": "ERNIE-5.0 (PayGo)", "model": "ernie-5.0", "tooltip": "文心新一代模型文心5.0是原生全模态大模型，采用原生的全模态统一建模技术，将文本、图像、音频、视频联合建模，具备综合的全模态能力。文心5.0基础能力全面升级，基准测试集表现出色，多模态理解、指令遵循、创意写作、事实性、智能体规划与工具应用等表现尤其出色。", "maxInputTokens": 112000, "maxOutputTokens": 16000, "thinking": ["enabled", "disabled"], "capabilities": { "toolCalling": true, "imageInput": true }, "tokenPricing": { "pricing": { "RMB": [6, 24] }, "tiers": [{ "contextSizeMin": 32001, "pricing": { "RMB": [10, 40] } }] } }, { "id": "ernie-5.1", "name": "ERNIE-5.1 (PayGo)", "model": "ernie-5.1", "tooltip": "ERNIE 5.1是文心系列最新模型，基础能力全面升级，在智能体、知识、推理、深度搜索等方面均有显著提升。此次升级发布，采用分离式全异步强化学习技术，针对性解决大模型向智能体自主决策演进中面临的训推数值偏差、异构资源利用率低及长尾效应导致的全局等挑战。配套采用规模化智能体后训练技术来提升模型能力与泛化表现，让环境、专家、融合三步协同 ，既保证了训练效率，也显著提升模型在复杂任务中的稳定性与表现。", "maxInputTokens": 112000, "maxOutputTokens": 16000, "thinking": ["enabled", "disabled"], "capabilities": { "toolCalling": true, "imageInput": false }, "tokenPricing": { "pricing": { "RMB": [4, 18] }, "tiers": [{ "contextSizeMin": 32001, "pricing": { "RMB": [6, 22] } }] } } ] }