💄 style: Update doubao-seed-1.6-vision models (#9052)

*  feat: 更新 Hunyuan、Novita 和 VolcEngine 模型的参数和描述

*  feat: 添加 DeepSeek V3.1 的思考模式和非思考模式模型,更新描述和显示名称

*  feat: 更新 Groq 模型的描述和显示名称,调整 ID 格式

*  feat: 更新 Novita 模型的上下文窗口令牌和定价策略

*  feat: 添加 Gemma 3 12B 和 Seed OSS 36B Instruct 模型,更新描述、定价和上下文窗口令牌

*  feat: 更新 Novita 模型的定价策略,调整输入和输出的费率

*  feat: 移除 Gemini 2.5 Flash 实验模型,更新模型列表

*  feat: 添加 Kimi K2 0905 模型,更新上下文窗口令牌和定价策略

* update groq

*  feat: 添加 Kimi K2 0905 模型,更新上下文窗口令牌和定价策略

*  feat: 更新 Doubao Seed 模型的部署名称和最大输出,调整上下文窗口令牌

*  feat: 添加 Qwen3 Max Preview 模型,更新上下文窗口令牌和定价策略
This commit is contained in:
sxjeru
2025-09-06 11:30:46 +08:00
committed by GitHub
parent a66856dc83
commit df2d001336
10 changed files with 368 additions and 60 deletions
+38 -4
View File
@@ -552,13 +552,11 @@ const aihubmixModels: AIChatModelCard[] = [
{
abilities: {
functionCall: true,
reasoning: true,
},
contextWindowTokens: 131_072,
description:
'DeepSeek-V3.1 是深度求索全新推出的混合推理模型,支持思考与非思考2种推理模式,较 DeepSeek-R1-0528 思考效率更高。经 Post-Training 优化,Agent 工具使用与智能体任务表现大幅提升。',
displayName: 'DeepSeek V3.1',
enabled: true,
'DeepSeek-V3.1-非思考模式;DeepSeek-V3.1 是深度求索全新推出的混合推理模型,支持思考与非思考2种推理模式,较 DeepSeek-R1-0528 思考效率更高。经 Post-Training 优化,Agent 工具使用与智能体任务表现大幅提升。',
displayName: 'DeepSeek V3.1 (non-Think)',
id: 'DeepSeek-V3.1',
pricing: {
units: [
@@ -568,6 +566,42 @@ const aihubmixModels: AIChatModelCard[] = [
},
type: 'chat',
},
{
abilities: {
functionCall: true,
reasoning: true,
},
contextWindowTokens: 131_072,
description:
'DeepSeek-V3.1-思考模式;DeepSeek-V3.1 是深度求索全新推出的混合推理模型,支持思考与非思考2种推理模式,较 DeepSeek-R1-0528 思考效率更高。经 Post-Training 优化,Agent 工具使用与智能体任务表现大幅提升。',
displayName: 'DeepSeek V3.1 (Think)',
id: 'DeepSeek-V3.1-Think',
pricing: {
units: [
{ name: 'textInput', rate: 0.56, strategy: 'fixed', unit: 'millionTokens' },
{ name: 'textOutput', rate: 1.68, strategy: 'fixed', unit: 'millionTokens' },
],
},
type: 'chat',
},
{
abilities: {
functionCall: true,
reasoning: true,
},
contextWindowTokens: 131_072,
description:
'DeepSeek V3.1 Fast 是 DeepSeek V3.1版本的高TPS极速版。 混合思考模式:通过更改聊天模板,一个模型可以同时支持思考模式和非思考模式。 更智能的工具调用:通过后训练优化,模型在工具使用和代理任务中的表现显著提升。',
displayName: 'DeepSeek V3.1 (Fast)',
id: 'DeepSeek-V3.1-Fast',
pricing: {
units: [
{ name: 'textInput', rate: 1.096, strategy: 'fixed', unit: 'millionTokens' },
{ name: 'textOutput', rate: 3.288, strategy: 'fixed', unit: 'millionTokens' },
],
},
type: 'chat',
},
{
abilities: {
functionCall: true,
+26 -8
View File
@@ -7,19 +7,19 @@ const groqChatModels: AIChatModelCard[] = [
{
contextWindowTokens: 131_072,
description:
'Compound-beta 是一个复合 AI 系统,由 GroqCloud 中已经支持的多个开放可用的模型提供支持,可以智能地、有选择地使用工具来回答用户查询。',
displayName: 'Compound Beta',
'Compound 是一个复合 AI 系统,由 GroqCloud 中已经支持的多个开放可用的模型提供支持,可以智能地、有选择地使用工具来回答用户查询。',
displayName: 'Compound',
enabled: true,
id: 'compound-beta',
id: 'groq/compound',
maxOutput: 8192,
type: 'chat',
},
{
contextWindowTokens: 131_072,
description:
'Compound-beta-mini 是一个复合 AI 系统,由 GroqCloud 中已经支持的公开可用模型提供支持,可以智能地、有选择地使用工具来回答用户查询。',
displayName: 'Compound Beta Mini',
id: 'compound-beta-mini',
'Compound-mini 是一个复合 AI 系统,由 GroqCloud 中已经支持的公开可用模型提供支持,可以智能地、有选择地使用工具来回答用户查询。',
displayName: 'Compound Mini',
id: 'groq/compound-mini',
maxOutput: 8192,
type: 'chat',
},
@@ -63,6 +63,25 @@ const groqChatModels: AIChatModelCard[] = [
releasedAt: '2025-08-06',
type: 'chat',
},
{
abilities: {
functionCall: true,
},
contextWindowTokens: 262_144,
description:
'kimi-k2-0905-preview 模型上下文长度为 256k,具备更强的 Agentic Coding 能力、更突出的前端代码的美观度和实用性、以及更好的上下文理解能力。',
displayName: 'Kimi K2 0905',
enabled: true,
id: 'moonshotai/kimi-k2-instruct-0905',
pricing: {
units: [
{ name: 'textInput', rate: 1, strategy: 'fixed', unit: 'millionTokens' },
{ name: 'textOutput', rate: 3, strategy: 'fixed', unit: 'millionTokens' },
],
},
releasedAt: '2025-09-05',
type: 'chat',
},
{
abilities: {
functionCall: true,
@@ -70,8 +89,7 @@ const groqChatModels: AIChatModelCard[] = [
contextWindowTokens: 131_072,
description:
'kimi-k2 是一款具备超强代码和 Agent 能力的 MoE 架构基础模型,总参数 1T,激活参数 32B。在通用知识推理、编程、数学、Agent 等主要类别的基准性能测试中,K2 模型的性能超过其他主流开源模型。',
displayName: 'Kimi K2 Instruct',
enabled: true,
displayName: 'Kimi K2 0711',
id: 'moonshotai/kimi-k2-instruct',
maxOutput: 16_384,
pricing: {
+3 -3
View File
@@ -26,9 +26,9 @@ const hunyuanChatModels: AIChatModelCard[] = [
reasoning: true,
search: true,
},
contextWindowTokens: 92_000,
contextWindowTokens: 96_000,
description:
'业内首个超大规模 Hybrid-Transformer-Mamba 推理模型,扩展推理能力,超强解码速度,进一步对齐人类偏好。',
'大幅提升主模型慢思考模型的高难数学、复杂推理、高难代码、指令遵循、文本创作质量等能力。',
displayName: 'Hunyuan T1',
enabled: true,
id: 'hunyuan-t1-latest',
@@ -40,7 +40,7 @@ const hunyuanChatModels: AIChatModelCard[] = [
{ name: 'textOutput', rate: 4, strategy: 'fixed', unit: 'millionTokens' },
],
},
releasedAt: '2025-05-21',
releasedAt: '2025-08-22',
settings: {
searchImpl: 'params',
},
+13 -2
View File
@@ -1,6 +1,19 @@
import { AIChatModelCard } from '../types/aiModel';
const modelscopeChatModels: AIChatModelCard[] = [
{
abilities: {
functionCall: true,
},
contextWindowTokens: 262_144,
description:
'kimi-k2-0905-preview 模型上下文长度为 256k,具备更强的 Agentic Coding 能力、更突出的前端代码的美观度和实用性、以及更好的上下文理解能力。',
displayName: 'Kimi K2 0905',
enabled: true,
id: 'moonshotai/Kimi-K2-Instruct-0905',
releasedAt: '2025-09-05',
type: 'chat',
},
{
abilities: {
functionCall: true,
@@ -53,7 +66,6 @@ const modelscopeChatModels: AIChatModelCard[] = [
contextWindowTokens: 131_072,
description: 'Qwen3-235B-A22B是通义千问3代超大规模模型,提供顶级的AI能力。',
displayName: 'Qwen3-235B-A22B',
enabled: true,
id: 'Qwen/Qwen3-235B-A22B',
type: 'chat',
},
@@ -64,7 +76,6 @@ const modelscopeChatModels: AIChatModelCard[] = [
contextWindowTokens: 131_072,
description: 'Qwen3-32B是通义千问3代模型,具有强大的推理和对话能力。',
displayName: 'Qwen3-32B',
enabled: true,
id: 'Qwen/Qwen3-32B',
type: 'chat',
},
+25 -5
View File
@@ -2,6 +2,27 @@ import { AIChatModelCard } from '../types/aiModel';
// https://platform.moonshot.cn/docs/pricing/chat
const moonshotChatModels: AIChatModelCard[] = [
{
abilities: {
functionCall: true,
},
contextWindowTokens: 262_144,
description:
'kimi-k2-0905-preview 模型上下文长度为 256k,具备更强的 Agentic Coding 能力、更突出的前端代码的美观度和实用性、以及更好的上下文理解能力。',
displayName: 'Kimi K2 0905',
enabled: true,
id: 'kimi-k2-0905-preview',
pricing: {
currency: 'CNY',
units: [
{ name: 'textInput_cacheRead', rate: 1, strategy: 'fixed', unit: 'millionTokens' },
{ name: 'textInput', rate: 4, strategy: 'fixed', unit: 'millionTokens' },
{ name: 'textOutput', rate: 16, strategy: 'fixed', unit: 'millionTokens' },
],
},
releasedAt: '2025-09-05',
type: 'chat',
},
{
abilities: {
functionCall: true,
@@ -9,8 +30,7 @@ const moonshotChatModels: AIChatModelCard[] = [
contextWindowTokens: 131_072,
description:
'kimi-k2 是一款具备超强代码和 Agent 能力的 MoE 架构基础模型,总参数 1T,激活参数 32B。在通用知识推理、编程、数学、Agent 等主要类别的基准性能测试中,K2 模型的性能超过其他主流开源模型。',
displayName: 'Kimi K2',
enabled: true,
displayName: 'Kimi K2 0711',
id: 'kimi-k2-0711-preview',
pricing: {
currency: 'CNY',
@@ -27,10 +47,10 @@ const moonshotChatModels: AIChatModelCard[] = [
abilities: {
functionCall: true,
},
contextWindowTokens: 131_072,
contextWindowTokens: 262_144,
description:
'kimi-k2 是一款具备超强代码和 Agent 能力的 MoE 架构基础模型,总参数 1T,激活参数 32B。在通用知识推理、编程、数学、Agent 等主要类别的基准性能测试中,K2 模型的性能超过其他主流开源模型。',
displayName: 'Kimi K2 Turbo',
displayName: 'Kimi K2 0905 Turbo',
id: 'kimi-k2-turbo-preview',
pricing: {
currency: 'CNY',
@@ -40,7 +60,7 @@ const moonshotChatModels: AIChatModelCard[] = [
{ name: 'textOutput', rate: 64, strategy: 'fixed', unit: 'millionTokens' },
],
},
releasedAt: '2025-07-11',
releasedAt: '2025-09-05',
type: 'chat',
},
{
+40 -9
View File
@@ -2,6 +2,24 @@ import { AIChatModelCard } from '../types/aiModel';
// https://novita.ai/pricing
const novitaChatModels: AIChatModelCard[] = [
{
abilities: {
functionCall: true,
},
contextWindowTokens: 262_144,
description:
'kimi-k2-0905-preview 模型上下文长度为 256k,具备更强的 Agentic Coding 能力、更突出的前端代码的美观度和实用性、以及更好的上下文理解能力。',
displayName: 'Kimi K2 0905',
id: 'moonshotai/kimi-k2-0905',
pricing: {
units: [
{ name: 'textInput', rate: 0.6, strategy: 'fixed', unit: 'millionTokens' },
{ name: 'textOutput', rate: 2.5, strategy: 'fixed', unit: 'millionTokens' },
],
},
releasedAt: '2025-09-05',
type: 'chat',
},
{
abilities: {
functionCall: true,
@@ -9,11 +27,12 @@ const novitaChatModels: AIChatModelCard[] = [
},
contextWindowTokens: 163_840,
displayName: 'DeepSeek V3.1',
enabled: true,
id: 'deepseek/deepseek-v3.1',
pricing: {
units: [
{ name: 'textInput', rate: 0.55, strategy: 'fixed', unit: 'millionTokens' },
{ name: 'textOutput', rate: 1.66, strategy: 'fixed', unit: 'millionTokens' },
{ name: 'textInput', rate: 0.27, strategy: 'fixed', unit: 'millionTokens' },
{ name: 'textOutput', rate: 1, strategy: 'fixed', unit: 'millionTokens' },
],
},
type: 'chat',
@@ -128,10 +147,10 @@ const novitaChatModels: AIChatModelCard[] = [
},
{
abilities: {
reasoning: true,
functionCall: true,
},
contextWindowTokens: 131_072,
displayName: 'Kimi K2 Instruct',
displayName: 'Kimi K2 0711',
id: 'moonshotai/kimi-k2-instruct',
pricing: {
units: [
@@ -286,12 +305,12 @@ const novitaChatModels: AIChatModelCard[] = [
abilities: {
reasoning: true,
},
contextWindowTokens: 40_960,
contextWindowTokens: 32_768,
displayName: 'Qwen3 30B A3B FP8',
id: 'qwen/qwen3-30b-a3b-fp8',
pricing: {
units: [
{ name: 'textInput', rate: 0.1, strategy: 'fixed', unit: 'millionTokens' },
{ name: 'textInput', rate: 0.09, strategy: 'fixed', unit: 'millionTokens' },
{ name: 'textOutput', rate: 0.45, strategy: 'fixed', unit: 'millionTokens' },
],
},
@@ -413,7 +432,7 @@ const novitaChatModels: AIChatModelCard[] = [
type: 'chat',
},
{
contextWindowTokens: 32_000,
contextWindowTokens: 32_768,
description: 'Gemma 3 27B 是谷歌的一款开源语言模型,以其在效率和性能方面设立了新的标准。',
displayName: 'Gemma 3 27B',
id: 'google/gemma-3-27b-it',
@@ -425,6 +444,20 @@ const novitaChatModels: AIChatModelCard[] = [
},
type: 'chat',
},
{
contextWindowTokens: 131_072,
description: 'Gemma 3 12B 是谷歌的一款开源语言模型,以其在效率和性能方面设立了新的标准。',
displayName: 'Gemma 3 12B',
id: 'google/gemma-3-12b-it',
maxOutput: 8192,
pricing: {
units: [
{ name: 'textInput', rate: 0.05, strategy: 'fixed', unit: 'millionTokens' },
{ name: 'textOutput', rate: 0.1, strategy: 'fixed', unit: 'millionTokens' },
],
},
type: 'chat',
},
{
contextWindowTokens: 32_768,
description: 'Gemma 3 1B 是谷歌的一款开源语言模型,以其在效率和性能方面设立了新的标准。',
@@ -549,7 +582,6 @@ const novitaChatModels: AIChatModelCard[] = [
},
contextWindowTokens: 163_840,
displayName: 'Deepseek V3 0324',
enabled: true,
id: 'deepseek/deepseek-v3-0324',
pricing: {
units: [
@@ -566,7 +598,6 @@ const novitaChatModels: AIChatModelCard[] = [
},
contextWindowTokens: 163_840,
displayName: 'Deepseek R1 0528',
enabled: true,
id: 'deepseek/deepseek-r1-0528',
pricing: {
units: [
@@ -50,19 +50,6 @@ const openrouterChatModels: AIChatModelCard[] = [
releasedAt: '2025-08-26',
type: 'chat',
},
{
abilities: {
imageOutput: true,
vision: true,
},
contextWindowTokens: 32_768 + 8192,
description: 'Gemini 2.5 Flash 实验模型,支持图像生成',
displayName: 'Nano Banana (free)',
id: 'google/gemini-2.5-flash-image-preview:free',
maxOutput: 8192,
releasedAt: '2025-08-26',
type: 'chat',
},
{
abilities: {
functionCall: true,
+62 -1
View File
@@ -601,6 +601,68 @@ const qwenChatModels: AIChatModelCard[] = [
},
type: 'chat',
},
{
abilities: {
functionCall: true,
},
config: {
deploymentName: 'qwen3-max-preview',
},
contextWindowTokens: 262_144,
description:
'通义千问3系列Max模型Preview版本,相较2.5系列整体通用能力有大幅度提升,中英文通用文本理解能力、复杂指令遵循能力、主观开放任务能力、多语言能力、工具调用能力均显著增强;模型知识幻觉更少。',
displayName: 'Qwen3 Max Preview',
enabled: true,
id: 'qwen3-max-preview',
maxOutput: 32_768,
organization: 'Qwen',
pricing: {
currency: 'CNY',
units: [
{
lookup: {
prices: {
'[0, 32_000]': 6 * 0.2,
'[32_000, 128_000]': 10 * 0.2,
'[128_000, infinity]': 15 * 0.2,
},
pricingParams: ['textInputRange'],
},
name: 'textInput_cacheRead',
strategy: 'lookup',
unit: 'millionTokens',
},
{
lookup: {
prices: {
'[0, 32_000]': 6,
'[32_000, 128_000]': 10,
'[128_000, infinity]': 15,
},
pricingParams: ['textInputRange'],
},
name: 'textInput',
strategy: 'lookup',
unit: 'millionTokens',
},
{
lookup: {
prices: {
'[0, 32_000]': 24,
'[32_000, 128_000]': 40,
'[128_000, infinity]': 60,
},
pricingParams: ['textInputRange'],
},
name: 'textOutput',
strategy: 'lookup',
unit: 'millionTokens',
},
],
},
releasedAt: '2025-09-05',
type: 'chat',
},
{
abilities: {
functionCall: true,
@@ -613,7 +675,6 @@ const qwenChatModels: AIChatModelCard[] = [
description:
'通义千问千亿级别超大规模语言模型,支持中文、英文等不同语言输入,当前通义千问2.5产品版本背后的API模型。',
displayName: 'Qwen Max',
enabled: true,
id: 'qwen-max',
maxOutput: 8192,
organization: 'Qwen',
@@ -47,6 +47,26 @@ const siliconcloudChatModels: AIChatModelCard[] = [
},
type: 'chat',
},
{
abilities: {
functionCall: true,
reasoning: true,
},
contextWindowTokens: 256_000,
description:
'Seed-OSS 是由字节跳动 Seed 团队开发的一系列开源大型语言模型,专为强大的长上下文处理、推理、智能体(agent)和通用能力而设计。该系列中的 Seed-OSS-36B-Instruct 是一个拥有 360 亿参数的指令微调模型,它原生支持超长上下文长度,使其能够一次性处理海量文档或复杂的代码库。该模型在推理、代码生成和智能体任务(如工具使用)方面进行了特别优化,同时保持了平衡且出色的通用能力。此模型的一大特色是“思考预算”(Thinking Budget)功能,允许用户根据需要灵活调整推理长度,从而在实际应用中有效提升推理效率。',
displayName: 'Seed OSS 36B Instruct',
id: 'ByteDance-Seed/Seed-OSS-36B-Instruct',
pricing: {
currency: 'CNY',
units: [
{ name: 'textInput', rate: 1.5, strategy: 'fixed', unit: 'millionTokens' },
{ name: 'textOutput', rate: 4, strategy: 'fixed', unit: 'millionTokens' },
],
},
releasedAt: '2025-08-20',
type: 'chat',
},
{
abilities: {
functionCall: true,
+141 -15
View File
@@ -1,7 +1,6 @@
import { AIChatModelCard, AIImageModelCard } from '../types/aiModel';
// modelInfo https://www.volcengine.com/docs/82379/1330310
// pricing https://console.volcengine.com/ark/region:ark+cn-beijing/openManagement
// https://www.volcengine.com/docs/82379/1330310
const doubaoChatModels: AIChatModelCard[] = [
{
@@ -42,7 +41,6 @@ const doubaoChatModels: AIChatModelCard[] = [
description:
'Kimi-K2 是一款Moonshot AI推出的具备超强代码和 Agent 能力的 MoE 架构基础模型,总参数 1T,激活参数 32B。在通用知识推理、编程、数学、Agent 等主要类别的基准性能测试中,K2 模型的性能超过其他主流开源模型。',
displayName: 'Kimi K2',
enabled: true,
id: 'kimi-k2',
maxOutput: 16_384,
pricing: {
@@ -61,7 +59,59 @@ const doubaoChatModels: AIChatModelCard[] = [
vision: true,
},
config: {
deploymentName: 'doubao-seed-1-6-thinking-250615',
deploymentName: 'doubao-seed-1-6-vision-250815',
},
contextWindowTokens: 256_000,
description:
'Doubao-Seed-1.6-vision 视觉深度思考模型,在教育、图像审核、巡检与安防和AI 搜索问答等场景下展现出更强的通用多模态理解和推理能力。支持 256k 上下文窗口,输出长度支持最大 64k tokens。',
displayName: 'Doubao Seed 1.6 Vision',
id: 'doubao-seed-1.6-vision',
maxOutput: 32_000,
pricing: {
currency: 'CNY',
units: [
{
lookup: {
prices: {
'[0, 32_000]': 0.8,
'[32_000, 128_000]': 2.4,
'[128_000, infinity]': 4.8,
},
pricingParams: ['textInputRange'],
},
name: 'textInput',
strategy: 'lookup',
unit: 'millionTokens',
},
{
lookup: {
prices: {
'[0, 32_000]': 8,
'[32_000, 128_000]': 16,
'[128_000, infinity]': 24,
},
pricingParams: ['textInputRange'],
},
name: 'textOutput',
strategy: 'lookup',
unit: 'millionTokens',
},
{ name: 'textInput_cacheRead', rate: 0.16, strategy: 'fixed', unit: 'millionTokens' },
],
},
settings: {
extendParams: ['enableReasoning'],
},
type: 'chat',
},
{
abilities: {
functionCall: true,
reasoning: true,
vision: true,
},
config: {
deploymentName: 'doubao-seed-1-6-thinking-250715',
},
contextWindowTokens: 256_000,
description:
@@ -69,12 +119,37 @@ const doubaoChatModels: AIChatModelCard[] = [
displayName: 'Doubao Seed 1.6 Thinking',
enabled: true,
id: 'doubao-seed-1.6-thinking',
maxOutput: 16_000,
maxOutput: 32_000,
pricing: {
currency: 'CNY',
units: [
{ name: 'textInput', rate: 1.2, strategy: 'fixed', unit: 'millionTokens' },
{ name: 'textOutput', rate: 16, strategy: 'fixed', unit: 'millionTokens' },
{
lookup: {
prices: {
'[0, 32_000]': 0.8,
'[32_000, 128_000]': 1.2,
'[128_000, infinity]': 2.4,
},
pricingParams: ['textInputRange'],
},
name: 'textInput',
strategy: 'lookup',
unit: 'millionTokens',
},
{
lookup: {
prices: {
'[0, 32_000]': 8,
'[32_000, 128_000]': 16,
'[128_000, infinity]': 24,
},
pricingParams: ['textInputRange'],
},
name: 'textOutput',
strategy: 'lookup',
unit: 'millionTokens',
},
{ name: 'textInput_cacheRead', rate: 0.16, strategy: 'fixed', unit: 'millionTokens' },
],
},
type: 'chat',
@@ -94,12 +169,38 @@ const doubaoChatModels: AIChatModelCard[] = [
displayName: 'Doubao Seed 1.6',
enabled: true,
id: 'doubao-seed-1.6',
maxOutput: 16_000,
maxOutput: 32_000,
pricing: {
currency: 'CNY',
units: [
{ name: 'textInput', rate: 1.2, strategy: 'fixed', unit: 'millionTokens' },
{ name: 'textOutput', rate: 16, strategy: 'fixed', unit: 'millionTokens' },
{
lookup: {
prices: {
'[0, 32_000]': 0.8,
'[32_000, 128_000]': 1.2,
'[128_000, infinity]': 2.4,
},
pricingParams: ['textInputRange'],
},
name: 'textInput',
strategy: 'lookup',
unit: 'millionTokens',
},
{
lookup: {
prices: {
'[0, 32_000]_[0, 8192]': 2,
'[0, 32_000]_[8192, infinity]': 8,
'[32_000, 128_000]_[0, infinity]': 16,
'[128_000, infinity]_[0, infinity]': 24,
},
pricingParams: ['textInputRange', 'textOutputRange'],
},
name: 'textOutput',
strategy: 'lookup',
unit: 'millionTokens',
},
{ name: 'textInput_cacheRead', rate: 0.16, strategy: 'fixed', unit: 'millionTokens' },
],
},
settings: {
@@ -114,7 +215,7 @@ const doubaoChatModels: AIChatModelCard[] = [
vision: true,
},
config: {
deploymentName: 'doubao-seed-1-6-flash-250615',
deploymentName: 'doubao-seed-1-6-flash-250828',
},
contextWindowTokens: 256_000,
description:
@@ -122,12 +223,37 @@ const doubaoChatModels: AIChatModelCard[] = [
displayName: 'Doubao Seed 1.6 Flash',
enabled: true,
id: 'doubao-seed-1.6-flash',
maxOutput: 16_000,
maxOutput: 32_000,
pricing: {
currency: 'CNY',
units: [
{ name: 'textInput', rate: 0.3, strategy: 'fixed', unit: 'millionTokens' },
{ name: 'textOutput', rate: 3, strategy: 'fixed', unit: 'millionTokens' },
{
lookup: {
prices: {
'[0, 32_000]': 0.15,
'[32_000, 128_000]': 0.3,
'[128_000, infinity]': 0.6,
},
pricingParams: ['textInputRange'],
},
name: 'textInput',
strategy: 'lookup',
unit: 'millionTokens',
},
{
lookup: {
prices: {
'[0, 32_000]': 1.5,
'[32_000, 128_000]': 3,
'[128_000, infinity]': 6,
},
pricingParams: ['textInputRange'],
},
name: 'textOutput',
strategy: 'lookup',
unit: 'millionTokens',
},
{ name: 'textInput_cacheRead', rate: 0.03, strategy: 'fixed', unit: 'millionTokens' },
],
},
settings: {
@@ -235,7 +361,7 @@ const doubaoChatModels: AIChatModelCard[] = [
],
},
settings: {
extendParams: ['enableReasoning'],
extendParams: ['thinking'],
},
type: 'chat',
},