Skip to content

Commit 02136f5

Browse files
authored
💄 style: Add Volcano Ark models (lobehub#6368)
* Update volcengine.ts * Update volcengine.ts * Update volcengine.ts * Update volcengine.ts * Update volcengine.ts * Update volcengine.ts
1 parent fbf9fb1 commit 02136f5

File tree

1 file changed

+249
-2
lines changed

1 file changed

+249
-2
lines changed

src/config/aiModels/volcengine.ts

Lines changed: 249 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,28 +1,216 @@
11
import { AIChatModelCard } from '@/types/aiModel';
22

3+
// modelInfo https://www.volcengine.com/docs/82379/1330310
4+
// pricing https://console.volcengine.com/ark/region:ark+cn-beijing/openManagement
5+
36
const doubaoChatModels: AIChatModelCard[] = [
47
{
58
abilities: {
69
reasoning: true,
710
},
11+
config: {
12+
deploymentName: 'deepseek-r1-250120',
13+
},
814
contextWindowTokens: 65_536,
915
description:
10-
'拥有极致的响应速度,更好的性价比,为客户不同场景提供更灵活的选择。支持 4k 上下文窗口的推理和精调。',
16+
'DeepSeek-R1 在后训练阶段大规模使用了强化学习技术,在仅有极少标注数据的情况下,极大提升了模型推理能力。在数学、代码、自然语言推理等任务上,性能比肩 OpenAI o1 正式版。',
1117
displayName: 'DeepSeek R1',
1218
enabled: true,
1319
id: 'deepseek-r1',
20+
maxOutput: 8000,
21+
pricing: {
22+
currency: 'CNY',
23+
input: 2, // 2.24 恢复原价 4, 16
24+
output: 8,
25+
},
26+
type: 'chat',
27+
},
28+
{
29+
abilities: {
30+
reasoning: true,
31+
},
32+
config: {
33+
deploymentName: 'deepseek-r1-distill-qwen-32b-250120',
34+
},
35+
contextWindowTokens: 32_768,
36+
description:
37+
'DeepSeek-R1-Distill 模型是在开源模型的基础上通过微调训练得到的,训练过程中使用了由 DeepSeek-R1 生成的样本数据。',
38+
displayName: 'DeepSeek R1 Distill Qwen 32B',
39+
id: 'deepseek-r1-distill-qwen-32b',
40+
maxOutput: 8000,
41+
pricing: {
42+
currency: 'CNY',
43+
input: 1.5,
44+
output: 6,
45+
},
46+
type: 'chat',
47+
},
48+
{
49+
abilities: {
50+
reasoning: true,
51+
},
52+
config: {
53+
deploymentName: 'deepseek-r1-distill-qwen-7b-250120',
54+
},
55+
contextWindowTokens: 32_768,
56+
description:
57+
'DeepSeek-R1-Distill 模型是在开源模型的基础上通过微调训练得到的,训练过程中使用了由 DeepSeek-R1 生成的样本数据。',
58+
displayName: 'DeepSeek R1 Distill Qwen 7B',
59+
id: 'deepseek-r1-distill-qwen-7b',
60+
maxOutput: 8000,
61+
pricing: {
62+
currency: 'CNY',
63+
input: 0.6,
64+
output: 2.4,
65+
},
1466
type: 'chat',
1567
},
1668
{
1769
abilities: {
1870
functionCall: true,
1971
},
72+
config: {
73+
deploymentName: 'deepseek-v3-241226',
74+
},
2075
contextWindowTokens: 65_536,
2176
description:
2277
'DeepSeek-V3 是一款由深度求索公司自研的MoE模型。DeepSeek-V3 多项评测成绩超越了 Qwen2.5-72B 和 Llama-3.1-405B 等其他开源模型,并在性能上和世界顶尖的闭源模型 GPT-4o 以及 Claude-3.5-Sonnet 不分伯仲。',
2378
displayName: 'DeepSeek V3',
2479
enabled: true,
2580
id: 'deepseek-v3',
81+
maxOutput: 8000,
82+
pricing: {
83+
currency: 'CNY',
84+
input: 2,
85+
output: 8,
86+
},
87+
type: 'chat',
88+
},
89+
{
90+
abilities: {
91+
functionCall: true,
92+
},
93+
config: {
94+
deploymentName: 'doubao-1-5-pro-32k-250115',
95+
},
96+
contextWindowTokens: 32_768,
97+
description:
98+
'Doubao-1.5-pro 全新一代主力模型,性能全面升级,在知识、代码、推理、等方面表现卓越。',
99+
displayName: 'Doubao 1.5 Pro 32k',
100+
enabled: true,
101+
id: 'doubao-1.5-pro-32k',
102+
maxOutput: 12_000,
103+
pricing: {
104+
currency: 'CNY',
105+
input: 0.8,
106+
output: 2,
107+
},
108+
type: 'chat',
109+
},
110+
{
111+
abilities: {
112+
functionCall: true,
113+
},
114+
config: {
115+
deploymentName: 'doubao-1-5-pro-256k-250115',
116+
},
117+
contextWindowTokens: 32_768,
118+
description:
119+
'Doubao-1.5-pro-256k 基于 Doubao-1.5-Pro 全面升级版,整体效果大幅提升 10%。支持 256k 上下文窗口的推理,输出长度支持最大 12k tokens。更高性能、更大窗口、超高性价比,适用于更广泛的应用场景。',
120+
displayName: 'Doubao 1.5 Pro 256k',
121+
enabled: true,
122+
id: 'doubao-1.5-pro-256k',
123+
maxOutput: 12_000,
124+
pricing: {
125+
currency: 'CNY',
126+
input: 5,
127+
output: 9,
128+
},
129+
type: 'chat',
130+
},
131+
{
132+
abilities: {
133+
functionCall: true,
134+
},
135+
config: {
136+
deploymentName: 'doubao-1-5-lite-32k-250115',
137+
},
138+
contextWindowTokens: 32_768,
139+
description:
140+
'Doubao-1.5-lite 全新一代轻量版模型,极致响应速度,效果与时延均达到全球一流水平。',
141+
displayName: 'Doubao 1.5 Lite 32k',
142+
enabled: true,
143+
id: 'doubao-1.5-lite-32k',
144+
maxOutput: 12_000,
145+
pricing: {
146+
currency: 'CNY',
147+
input: 0.3,
148+
output: 0.6,
149+
},
150+
type: 'chat',
151+
},
152+
{
153+
abilities: {
154+
vision: true,
155+
},
156+
config: {
157+
deploymentName: 'doubao-1-5-vision-pro-32k-250115',
158+
},
159+
contextWindowTokens: 32_768,
160+
description:
161+
'Doubao-1.5-vision-pro 全新升级的多模态大模型,支持任意分辨率和极端长宽比图像识别,增强视觉推理、文档识别、细节信息理解和指令遵循能力。',
162+
displayName: 'Doubao 1.5 Vision Pro 32k',
163+
enabled: true,
164+
id: 'Doubao-1.5-vision-pro-32k',
165+
maxOutput: 12_000,
166+
pricing: {
167+
currency: 'CNY',
168+
input: 3,
169+
output: 9,
170+
},
171+
releasedAt: '2025-01-15',
172+
type: 'chat',
173+
},
174+
{
175+
abilities: {
176+
vision: true,
177+
},
178+
config: {
179+
deploymentName: 'doubao-vision-pro-32k-241028',
180+
},
181+
contextWindowTokens: 32_768,
182+
description:
183+
'Doubao-vision 模型是豆包推出的多模态大模型,具备强大的图片理解与推理能力,以及精准的指令理解能力。模型在图像文本信息抽取、基于图像的推理任务上有展现出了强大的性能,能够应用于更复杂、更广泛的视觉问答任务。',
184+
displayName: 'Doubao Vision Pro 32k',
185+
id: 'Doubao-vision-pro-32k',
186+
maxOutput: 4096,
187+
pricing: {
188+
currency: 'CNY',
189+
input: 3,
190+
output: 9,
191+
},
192+
releasedAt: '2024-10-28',
193+
type: 'chat',
194+
},
195+
{
196+
abilities: {
197+
vision: true,
198+
},
199+
config: {
200+
deploymentName: 'doubao-vision-lite-32k-241015',
201+
},
202+
contextWindowTokens: 32_768,
203+
description:
204+
'Doubao-vision 模型是豆包推出的多模态大模型,具备强大的图片理解与推理能力,以及精准的指令理解能力。模型在图像文本信息抽取、基于图像的推理任务上有展现出了强大的性能,能够应用于更复杂、更广泛的视觉问答任务。',
205+
displayName: 'Doubao Vision Lite 32k',
206+
id: 'Doubao-vision-lite-32k',
207+
maxOutput: 4096,
208+
pricing: {
209+
currency: 'CNY',
210+
input: 1.5,
211+
output: 4.5,
212+
},
213+
releasedAt: '2024-10-15',
26214
type: 'chat',
27215
},
28216
{
@@ -31,22 +219,46 @@ const doubaoChatModels: AIChatModelCard[] = [
31219
'拥有极致的响应速度,更好的性价比,为客户不同场景提供更灵活的选择。支持 4k 上下文窗口的推理和精调。',
32220
displayName: 'Doubao Lite 4k',
33221
id: 'Doubao-lite-4k',
222+
maxOutput: 4096,
223+
pricing: {
224+
currency: 'CNY',
225+
input: 0.3,
226+
output: 0.6,
227+
},
34228
type: 'chat',
35229
},
36230
{
231+
config: {
232+
deploymentName: 'doubao-lite-32k-240828',
233+
},
37234
contextWindowTokens: 32_768,
38235
description:
39236
'拥有极致的响应速度,更好的性价比,为客户不同场景提供更灵活的选择。支持 32k 上下文窗口的推理和精调。',
40237
displayName: 'Doubao Lite 32k',
41238
id: 'Doubao-lite-32k',
239+
maxOutput: 4096,
240+
pricing: {
241+
currency: 'CNY',
242+
input: 0.3,
243+
output: 0.6,
244+
},
42245
type: 'chat',
43246
},
44247
{
248+
config: {
249+
deploymentName: 'doubao-lite-128k-240828',
250+
},
45251
contextWindowTokens: 128_000,
46252
description:
47253
'拥有极致的响应速度,更好的性价比,为客户不同场景提供更灵活的选择。支持 128k 上下文窗口的推理和精调。',
48254
displayName: 'Doubao Lite 128k',
49255
id: 'Doubao-lite-128k',
256+
maxOutput: 4096,
257+
pricing: {
258+
currency: 'CNY',
259+
input: 0.8,
260+
output: 1,
261+
},
50262
type: 'chat',
51263
},
52264
{
@@ -55,17 +267,29 @@ const doubaoChatModels: AIChatModelCard[] = [
55267
'效果最好的主力模型,适合处理复杂任务,在参考问答、总结摘要、创作、文本分类、角色扮演等场景都有很好的效果。支持 4k 上下文窗口的推理和精调。',
56268
displayName: 'Doubao Pro 4k',
57269
id: 'Doubao-pro-4k',
270+
maxOutput: 4096,
271+
pricing: {
272+
currency: 'CNY',
273+
input: 0.8,
274+
output: 2,
275+
},
58276
type: 'chat',
59277
},
60278
{
61279
config: {
62-
deploymentName: 'Doubao-pro-test',
280+
deploymentName: 'doubao-pro-32k-241215',
63281
},
64282
contextWindowTokens: 32_768,
65283
description:
66284
'效果最好的主力模型,适合处理复杂任务,在参考问答、总结摘要、创作、文本分类、角色扮演等场景都有很好的效果。支持 32k 上下文窗口的推理和精调。',
67285
displayName: 'Doubao Pro 32k',
68286
id: 'Doubao-pro-32k',
287+
maxOutput: 4096,
288+
pricing: {
289+
currency: 'CNY',
290+
input: 0.8,
291+
output: 2,
292+
},
69293
type: 'chat',
70294
},
71295
{
@@ -74,6 +298,29 @@ const doubaoChatModels: AIChatModelCard[] = [
74298
'效果最好的主力模型,适合处理复杂任务,在参考问答、总结摘要、创作、文本分类、角色扮演等场景都有很好的效果。支持 128k 上下文窗口的推理和精调。',
75299
displayName: 'Doubao Pro 128k',
76300
id: 'Doubao-pro-128k',
301+
maxOutput: 4096,
302+
pricing: {
303+
currency: 'CNY',
304+
input: 5,
305+
output: 9,
306+
},
307+
type: 'chat',
308+
},
309+
{
310+
config: {
311+
deploymentName: 'doubao-pro-256k-241115',
312+
},
313+
contextWindowTokens: 256_000,
314+
description:
315+
'效果最好的主力模型,适合处理复杂任务,在参考问答、总结摘要、创作、文本分类、角色扮演等场景都有很好的效果。支持 256k 上下文窗口的推理和精调。',
316+
displayName: 'Doubao Pro 256k',
317+
id: 'Doubao-pro-256k',
318+
maxOutput: 4096,
319+
pricing: {
320+
currency: 'CNY',
321+
input: 5,
322+
output: 9,
323+
},
77324
type: 'chat',
78325
},
79326
];

0 commit comments

Comments
 (0)