@lobehub/chat 1.136.11 → 1.136.12
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +26 -0
- package/changelog/v1.json +9 -0
- package/package.json +1 -1
- package/packages/model-bank/src/aiModels/infiniai.ts +465 -174
- package/packages/model-bank/src/aiModels/modelscope.ts +10 -20
- package/packages/model-bank/src/aiModels/novita.ts +2 -2
- package/packages/model-runtime/src/providers/infiniai/index.ts +8 -54
- package/src/app/[variants]/(main)/settings/provider/detail/bedrock/index.tsx +36 -2
- package/src/config/modelProviders/infiniai.ts +2 -25
- package/src/config/modelProviders/modelscope.ts +1 -17
package/CHANGELOG.md
CHANGED
|
@@ -2,6 +2,32 @@
|
|
|
2
2
|
|
|
3
3
|
# Changelog
|
|
4
4
|
|
|
5
|
+
### [Version 1.136.12](https://github.com/lobehub/lobe-chat/compare/v1.136.11...v1.136.12)
|
|
6
|
+
|
|
7
|
+
<sup>Released on **2025-10-11**</sup>
|
|
8
|
+
|
|
9
|
+
#### 💄 Styles
|
|
10
|
+
|
|
11
|
+
- **misc**: Add more AWS regions, Update infini-ai models.
|
|
12
|
+
|
|
13
|
+
<br/>
|
|
14
|
+
|
|
15
|
+
<details>
|
|
16
|
+
<summary><kbd>Improvements and Fixes</kbd></summary>
|
|
17
|
+
|
|
18
|
+
#### Styles
|
|
19
|
+
|
|
20
|
+
- **misc**: Add more AWS regions, closes [#9644](https://github.com/lobehub/lobe-chat/issues/9644) ([4a82daf](https://github.com/lobehub/lobe-chat/commit/4a82daf))
|
|
21
|
+
- **misc**: Update infini-ai models, closes [#9646](https://github.com/lobehub/lobe-chat/issues/9646) ([5274225](https://github.com/lobehub/lobe-chat/commit/5274225))
|
|
22
|
+
|
|
23
|
+
</details>
|
|
24
|
+
|
|
25
|
+
<div align="right">
|
|
26
|
+
|
|
27
|
+
[](#readme-top)
|
|
28
|
+
|
|
29
|
+
</div>
|
|
30
|
+
|
|
5
31
|
### [Version 1.136.11](https://github.com/lobehub/lobe-chat/compare/v1.136.10...v1.136.11)
|
|
6
32
|
|
|
7
33
|
<sup>Released on **2025-10-11**</sup>
|
package/changelog/v1.json
CHANGED
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@lobehub/chat",
|
|
3
|
-
"version": "1.136.
|
|
3
|
+
"version": "1.136.12",
|
|
4
4
|
"description": "Lobe Chat - an open-source, high-performance chatbot framework that supports speech synthesis, multimodal, and extensible Function Call plugin system. Supports one-click free deployment of your private ChatGPT/LLM web application.",
|
|
5
5
|
"keywords": [
|
|
6
6
|
"framework",
|
|
@@ -1,61 +1,70 @@
|
|
|
1
1
|
import { AIChatModelCard } from '../types/aiModel';
|
|
2
2
|
|
|
3
3
|
// https://cloud.infini-ai.com/genstudio/model
|
|
4
|
-
// All models are currently free
|
|
5
|
-
// Currently the platform doesn't support Function Call
|
|
6
4
|
|
|
7
5
|
const infiniaiChatModels: AIChatModelCard[] = [
|
|
8
6
|
{
|
|
9
7
|
abilities: {
|
|
8
|
+
functionCall: true,
|
|
10
9
|
reasoning: true,
|
|
11
10
|
},
|
|
12
|
-
contextWindowTokens:
|
|
13
|
-
description:
|
|
14
|
-
|
|
15
|
-
displayName: 'DeepSeek R1',
|
|
11
|
+
contextWindowTokens: 131_072,
|
|
12
|
+
description: 'GLM-4.6 是智谱AI推出的最新大语言模型,具备更强的推理和生成能力。',
|
|
13
|
+
displayName: 'GLM-4.6',
|
|
16
14
|
enabled: true,
|
|
17
|
-
id: '
|
|
15
|
+
id: 'glm-4.6',
|
|
16
|
+
maxOutput: 4096,
|
|
18
17
|
pricing: {
|
|
19
18
|
currency: 'CNY',
|
|
20
19
|
units: [
|
|
21
|
-
{ name: 'textInput', rate:
|
|
22
|
-
{ name: 'textOutput', rate:
|
|
20
|
+
{ name: 'textInput', rate: 4, strategy: 'fixed', unit: 'millionTokens' },
|
|
21
|
+
{ name: 'textOutput', rate: 16, strategy: 'fixed', unit: 'millionTokens' },
|
|
23
22
|
],
|
|
24
23
|
},
|
|
24
|
+
settings: {
|
|
25
|
+
extendParams: ['enableReasoning'],
|
|
26
|
+
},
|
|
25
27
|
type: 'chat',
|
|
26
28
|
},
|
|
27
29
|
{
|
|
28
|
-
|
|
30
|
+
abilities: {
|
|
31
|
+
functionCall: true,
|
|
32
|
+
reasoning: true,
|
|
33
|
+
},
|
|
34
|
+
contextWindowTokens: 131_072,
|
|
29
35
|
description:
|
|
30
|
-
'DeepSeek-V3-
|
|
31
|
-
displayName: 'DeepSeek V3
|
|
36
|
+
'DeepSeek-V3.2-Exp 是深度求索推出的实验性大语言模型,具有更强的推理和生成能力。',
|
|
37
|
+
displayName: 'DeepSeek V3.2 Exp',
|
|
32
38
|
enabled: true,
|
|
33
|
-
id: 'deepseek-v3',
|
|
39
|
+
id: 'deepseek-v3.2-exp',
|
|
40
|
+
maxOutput: 65_536,
|
|
34
41
|
pricing: {
|
|
35
42
|
currency: 'CNY',
|
|
36
43
|
units: [
|
|
37
|
-
{ name: 'textInput', rate:
|
|
38
|
-
{ name: 'textOutput', rate:
|
|
44
|
+
{ name: 'textInput', rate: 2, strategy: 'fixed', unit: 'millionTokens' },
|
|
45
|
+
{ name: 'textOutput', rate: 3, strategy: 'fixed', unit: 'millionTokens' },
|
|
39
46
|
],
|
|
40
47
|
},
|
|
48
|
+
settings: {
|
|
49
|
+
extendParams: ['enableReasoning'],
|
|
50
|
+
},
|
|
41
51
|
type: 'chat',
|
|
42
52
|
},
|
|
43
53
|
{
|
|
44
54
|
abilities: {
|
|
45
55
|
functionCall: true,
|
|
46
|
-
|
|
56
|
+
vision: true,
|
|
47
57
|
},
|
|
48
|
-
contextWindowTokens:
|
|
49
|
-
description:
|
|
50
|
-
|
|
51
|
-
|
|
52
|
-
|
|
53
|
-
id: 'qwen3-235b-a22b',
|
|
58
|
+
contextWindowTokens: 131_072,
|
|
59
|
+
description: 'Qwen3 VL 235B A22B Instruct 是通义千问推出的多模态模型,支持视觉理解和推理。',
|
|
60
|
+
displayName: 'Qwen3 VL 235B A22B Instruct',
|
|
61
|
+
id: 'qwen3-vl-235b-a22b-instruct',
|
|
62
|
+
maxOutput: 32_768,
|
|
54
63
|
pricing: {
|
|
55
64
|
currency: 'CNY',
|
|
56
65
|
units: [
|
|
57
|
-
{ name: 'textInput', rate:
|
|
58
|
-
{ name: 'textOutput', rate:
|
|
66
|
+
{ name: 'textInput', rate: 2, strategy: 'fixed', unit: 'millionTokens' },
|
|
67
|
+
{ name: 'textOutput', rate: 8, strategy: 'fixed', unit: 'millionTokens' },
|
|
59
68
|
],
|
|
60
69
|
},
|
|
61
70
|
type: 'chat',
|
|
@@ -64,18 +73,18 @@ const infiniaiChatModels: AIChatModelCard[] = [
|
|
|
64
73
|
abilities: {
|
|
65
74
|
functionCall: true,
|
|
66
75
|
reasoning: true,
|
|
76
|
+
vision: true,
|
|
67
77
|
},
|
|
68
|
-
contextWindowTokens:
|
|
69
|
-
description:
|
|
70
|
-
|
|
71
|
-
|
|
72
|
-
|
|
73
|
-
id: 'qwen3-30b-a3b',
|
|
78
|
+
contextWindowTokens: 131_072,
|
|
79
|
+
description: 'Qwen3 VL 235B A22B Thinking 是通义千问推出的多模态推理模型,支持视觉理解和推理。',
|
|
80
|
+
displayName: 'Qwen3 VL 235B A22B Thinking',
|
|
81
|
+
id: 'qwen3-vl-235b-a22b-thinking',
|
|
82
|
+
maxOutput: 32_768,
|
|
74
83
|
pricing: {
|
|
75
84
|
currency: 'CNY',
|
|
76
85
|
units: [
|
|
77
|
-
{ name: 'textInput', rate:
|
|
78
|
-
{ name: 'textOutput', rate:
|
|
86
|
+
{ name: 'textInput', rate: 2, strategy: 'fixed', unit: 'millionTokens' },
|
|
87
|
+
{ name: 'textOutput', rate: 20, strategy: 'fixed', unit: 'millionTokens' },
|
|
79
88
|
],
|
|
80
89
|
},
|
|
81
90
|
type: 'chat',
|
|
@@ -85,19 +94,22 @@ const infiniaiChatModels: AIChatModelCard[] = [
|
|
|
85
94
|
functionCall: true,
|
|
86
95
|
reasoning: true,
|
|
87
96
|
},
|
|
88
|
-
contextWindowTokens:
|
|
97
|
+
contextWindowTokens: 131_072,
|
|
89
98
|
description:
|
|
90
|
-
'
|
|
91
|
-
displayName: '
|
|
92
|
-
|
|
93
|
-
|
|
99
|
+
'DeepSeek-V3.1-Terminus 是深度求索推出的终端优化版本大语言模型,专为终端设备优化。',
|
|
100
|
+
displayName: 'DeepSeek V3.1 Terminus',
|
|
101
|
+
id: 'deepseek-v3.1-terminus',
|
|
102
|
+
maxOutput: 65_536,
|
|
94
103
|
pricing: {
|
|
95
104
|
currency: 'CNY',
|
|
96
105
|
units: [
|
|
97
|
-
{ name: 'textInput', rate:
|
|
98
|
-
{ name: 'textOutput', rate:
|
|
106
|
+
{ name: 'textInput', rate: 4, strategy: 'fixed', unit: 'millionTokens' },
|
|
107
|
+
{ name: 'textOutput', rate: 12, strategy: 'fixed', unit: 'millionTokens' },
|
|
99
108
|
],
|
|
100
109
|
},
|
|
110
|
+
settings: {
|
|
111
|
+
extendParams: ['enableReasoning'],
|
|
112
|
+
},
|
|
101
113
|
type: 'chat',
|
|
102
114
|
},
|
|
103
115
|
{
|
|
@@ -105,17 +117,17 @@ const infiniaiChatModels: AIChatModelCard[] = [
|
|
|
105
117
|
functionCall: true,
|
|
106
118
|
reasoning: true,
|
|
107
119
|
},
|
|
108
|
-
contextWindowTokens:
|
|
120
|
+
contextWindowTokens: 131_072,
|
|
109
121
|
description:
|
|
110
|
-
'Qwen3-
|
|
111
|
-
displayName: 'Qwen3
|
|
112
|
-
|
|
113
|
-
|
|
122
|
+
'基于 Qwen3 的新一代思考模式开源模型,相较上一版本(通义千问3-235B-A22B-Thinking-2507)指令遵循能力有提升、模型总结回复更加精简。',
|
|
123
|
+
displayName: 'Qwen3 Next 80B A3B Thinking',
|
|
124
|
+
id: 'qwen3-next-80b-a3b-thinking',
|
|
125
|
+
maxOutput: 32_768,
|
|
114
126
|
pricing: {
|
|
115
127
|
currency: 'CNY',
|
|
116
128
|
units: [
|
|
117
|
-
{ name: 'textInput', rate:
|
|
118
|
-
{ name: 'textOutput', rate:
|
|
129
|
+
{ name: 'textInput', rate: 1, strategy: 'fixed', unit: 'millionTokens' },
|
|
130
|
+
{ name: 'textOutput', rate: 10, strategy: 'fixed', unit: 'millionTokens' },
|
|
119
131
|
],
|
|
120
132
|
},
|
|
121
133
|
type: 'chat',
|
|
@@ -123,57 +135,109 @@ const infiniaiChatModels: AIChatModelCard[] = [
|
|
|
123
135
|
{
|
|
124
136
|
abilities: {
|
|
125
137
|
functionCall: true,
|
|
126
|
-
reasoning: true,
|
|
127
138
|
},
|
|
128
|
-
contextWindowTokens:
|
|
139
|
+
contextWindowTokens: 131_072,
|
|
129
140
|
description:
|
|
130
|
-
'Qwen3-
|
|
131
|
-
displayName: 'Qwen3
|
|
132
|
-
|
|
133
|
-
|
|
141
|
+
'基于 Qwen3 的新一代非思考模式开源模型,相较上一版本(通义千问3-235B-A22B-Instruct-2507)中文文本理解能力更佳、逻辑推理能力有增强、文本生成类任务表现更好。',
|
|
142
|
+
displayName: 'Qwen3 Next 80B A3B Instruct',
|
|
143
|
+
id: 'qwen3-next-80b-a3b-instruct',
|
|
144
|
+
maxOutput: 32_768,
|
|
134
145
|
pricing: {
|
|
135
146
|
currency: 'CNY',
|
|
136
147
|
units: [
|
|
137
|
-
{ name: 'textInput', rate:
|
|
138
|
-
{ name: 'textOutput', rate:
|
|
148
|
+
{ name: 'textInput', rate: 1, strategy: 'fixed', unit: 'millionTokens' },
|
|
149
|
+
{ name: 'textOutput', rate: 10, strategy: 'fixed', unit: 'millionTokens' },
|
|
139
150
|
],
|
|
140
151
|
},
|
|
141
152
|
type: 'chat',
|
|
142
153
|
},
|
|
143
154
|
{
|
|
144
155
|
abilities: {
|
|
145
|
-
|
|
156
|
+
reasoning: true,
|
|
146
157
|
},
|
|
147
|
-
contextWindowTokens:
|
|
148
|
-
description:
|
|
149
|
-
|
|
150
|
-
|
|
151
|
-
|
|
152
|
-
id: 'qwen2.5-vl-72b-instruct',
|
|
158
|
+
contextWindowTokens: 131_072,
|
|
159
|
+
description: 'DeepSeek V3.1 模型为混合推理架构模型,同时支持思考模式与非思考模式。',
|
|
160
|
+
displayName: 'DeepSeek V3.1',
|
|
161
|
+
id: 'deepseek-v3.1',
|
|
162
|
+
maxOutput: 32_768,
|
|
153
163
|
pricing: {
|
|
154
164
|
currency: 'CNY',
|
|
155
165
|
units: [
|
|
156
|
-
{ name: 'textInput', rate:
|
|
157
|
-
{ name: 'textOutput', rate:
|
|
166
|
+
{ name: 'textInput', rate: 4, strategy: 'fixed', unit: 'millionTokens' },
|
|
167
|
+
{ name: 'textOutput', rate: 12, strategy: 'fixed', unit: 'millionTokens' },
|
|
158
168
|
],
|
|
159
169
|
},
|
|
160
170
|
type: 'chat',
|
|
161
171
|
},
|
|
162
172
|
{
|
|
163
173
|
abilities: {
|
|
174
|
+
reasoning: true,
|
|
175
|
+
},
|
|
176
|
+
contextWindowTokens: 64_000,
|
|
177
|
+
description: 'Baichuan M2 32B 是百川智能推出的混合专家模型,具备强大的推理能力。',
|
|
178
|
+
displayName: 'Baichuan M2 32B',
|
|
179
|
+
id: 'baichuan-m2-32b',
|
|
180
|
+
maxOutput: 8192,
|
|
181
|
+
pricing: {
|
|
182
|
+
currency: 'CNY',
|
|
183
|
+
units: [
|
|
184
|
+
{ name: 'textInput', rate: 2.9, strategy: 'fixed', unit: 'millionTokens' },
|
|
185
|
+
{ name: 'textOutput', rate: 11.6, strategy: 'fixed', unit: 'millionTokens' },
|
|
186
|
+
],
|
|
187
|
+
},
|
|
188
|
+
type: 'chat',
|
|
189
|
+
},
|
|
190
|
+
{
|
|
191
|
+
abilities: {
|
|
192
|
+
functionCall: true,
|
|
164
193
|
vision: true,
|
|
165
194
|
},
|
|
166
|
-
contextWindowTokens:
|
|
167
|
-
description:
|
|
168
|
-
|
|
169
|
-
|
|
170
|
-
|
|
171
|
-
id: 'qwen2.5-vl-32b-instruct',
|
|
195
|
+
contextWindowTokens: 131_072,
|
|
196
|
+
description: 'GLM-4.5V 是智谱AI推出的多模态模型,支持视觉理解和推理。',
|
|
197
|
+
displayName: 'GLM-4.5V',
|
|
198
|
+
id: 'glm-4.5v',
|
|
199
|
+
maxOutput: 4096,
|
|
172
200
|
pricing: {
|
|
173
201
|
currency: 'CNY',
|
|
174
202
|
units: [
|
|
175
|
-
{ name: 'textInput', rate:
|
|
176
|
-
{ name: 'textOutput', rate:
|
|
203
|
+
{ name: 'textInput', rate: 4, strategy: 'fixed', unit: 'millionTokens' },
|
|
204
|
+
{ name: 'textOutput', rate: 12, strategy: 'fixed', unit: 'millionTokens' },
|
|
205
|
+
],
|
|
206
|
+
},
|
|
207
|
+
type: 'chat',
|
|
208
|
+
},
|
|
209
|
+
{
|
|
210
|
+
abilities: {
|
|
211
|
+
reasoning: true,
|
|
212
|
+
},
|
|
213
|
+
contextWindowTokens: 131_072,
|
|
214
|
+
description: 'GLM-4.5系列模型是智谱AI专为智能体设计的混合推理模型,提供思考与非思考两种模式。',
|
|
215
|
+
displayName: 'GLM-4.5',
|
|
216
|
+
id: 'glm-4.5',
|
|
217
|
+
maxOutput: 4096,
|
|
218
|
+
pricing: {
|
|
219
|
+
currency: 'CNY',
|
|
220
|
+
units: [
|
|
221
|
+
{ name: 'textInput', rate: 2, strategy: 'fixed', unit: 'millionTokens' },
|
|
222
|
+
{ name: 'textOutput', rate: 8, strategy: 'fixed', unit: 'millionTokens' },
|
|
223
|
+
],
|
|
224
|
+
},
|
|
225
|
+
settings: {
|
|
226
|
+
extendParams: ['enableReasoning'],
|
|
227
|
+
},
|
|
228
|
+
type: 'chat',
|
|
229
|
+
},
|
|
230
|
+
{
|
|
231
|
+
contextWindowTokens: 131_072,
|
|
232
|
+
description: 'GLM-4.5-Air 是智谱AI推出的轻量级大语言模型,具备高效的推理能力。',
|
|
233
|
+
displayName: 'GLM-4.5-Air',
|
|
234
|
+
id: 'glm-4.5-air',
|
|
235
|
+
maxOutput: 4096,
|
|
236
|
+
pricing: {
|
|
237
|
+
currency: 'CNY',
|
|
238
|
+
units: [
|
|
239
|
+
{ name: 'textInput', rate: 0.6, strategy: 'fixed', unit: 'millionTokens' },
|
|
240
|
+
{ name: 'textOutput', rate: 4, strategy: 'fixed', unit: 'millionTokens' },
|
|
177
241
|
],
|
|
178
242
|
},
|
|
179
243
|
type: 'chat',
|
|
@@ -182,16 +246,35 @@ const infiniaiChatModels: AIChatModelCard[] = [
|
|
|
182
246
|
abilities: {
|
|
183
247
|
vision: true,
|
|
184
248
|
},
|
|
185
|
-
contextWindowTokens:
|
|
249
|
+
contextWindowTokens: 131_072,
|
|
250
|
+
description: 'Step3 是阶跃星辰推出的多模态模型,具备强大的视觉理解能力。',
|
|
251
|
+
displayName: 'Step3',
|
|
252
|
+
id: 'step3',
|
|
253
|
+
maxOutput: 4096,
|
|
254
|
+
pricing: {
|
|
255
|
+
currency: 'CNY',
|
|
256
|
+
units: [
|
|
257
|
+
{ name: 'textInput', rate: 1.5, strategy: 'fixed', unit: 'millionTokens' },
|
|
258
|
+
{ name: 'textOutput', rate: 4, strategy: 'fixed', unit: 'millionTokens' },
|
|
259
|
+
],
|
|
260
|
+
},
|
|
261
|
+
type: 'chat',
|
|
262
|
+
},
|
|
263
|
+
{
|
|
264
|
+
abilities: {
|
|
265
|
+
functionCall: true,
|
|
266
|
+
},
|
|
267
|
+
contextWindowTokens: 262_144,
|
|
186
268
|
description:
|
|
187
|
-
'
|
|
188
|
-
displayName: '
|
|
189
|
-
id: '
|
|
269
|
+
'通义千问代码模型开源版。最新的 qwen3-coder-480b-a35b-instruct 是基于 Qwen3 的代码生成模型,具有强大的Coding Agent能力,擅长工具调用和环境交互,能够实现自主编程、代码能力卓越的同时兼具通用能力。',
|
|
270
|
+
displayName: 'Qwen3 Coder 480B A35B',
|
|
271
|
+
id: 'qwen3-coder-480b-a35b-instruct',
|
|
272
|
+
maxOutput: 32_768,
|
|
190
273
|
pricing: {
|
|
191
274
|
currency: 'CNY',
|
|
192
275
|
units: [
|
|
193
|
-
{ name: 'textInput', rate:
|
|
194
|
-
{ name: 'textOutput', rate:
|
|
276
|
+
{ name: 'textInput', rate: 9, strategy: 'fixed', unit: 'millionTokens' },
|
|
277
|
+
{ name: 'textOutput', rate: 36, strategy: 'fixed', unit: 'millionTokens' },
|
|
195
278
|
],
|
|
196
279
|
},
|
|
197
280
|
type: 'chat',
|
|
@@ -200,17 +283,17 @@ const infiniaiChatModels: AIChatModelCard[] = [
|
|
|
200
283
|
abilities: {
|
|
201
284
|
functionCall: true,
|
|
202
285
|
},
|
|
203
|
-
contextWindowTokens:
|
|
286
|
+
contextWindowTokens: 131_072,
|
|
204
287
|
description:
|
|
205
|
-
'
|
|
206
|
-
displayName: '
|
|
207
|
-
|
|
208
|
-
|
|
288
|
+
'基于Qwen3的非思考模式开源模型,相较上一版本(通义千问3-235B-A22B)主观创作能力与模型安全性均有小幅度提升。',
|
|
289
|
+
displayName: 'Qwen3 235B A22B Instruct 2507',
|
|
290
|
+
id: 'qwen3-235b-a22b-instruct-2507',
|
|
291
|
+
maxOutput: 8192,
|
|
209
292
|
pricing: {
|
|
210
293
|
currency: 'CNY',
|
|
211
294
|
units: [
|
|
212
|
-
{ name: 'textInput', rate:
|
|
213
|
-
{ name: 'textOutput', rate:
|
|
295
|
+
{ name: 'textInput', rate: 2, strategy: 'fixed', unit: 'millionTokens' },
|
|
296
|
+
{ name: 'textOutput', rate: 8, strategy: 'fixed', unit: 'millionTokens' },
|
|
214
297
|
],
|
|
215
298
|
},
|
|
216
299
|
type: 'chat',
|
|
@@ -219,47 +302,109 @@ const infiniaiChatModels: AIChatModelCard[] = [
|
|
|
219
302
|
abilities: {
|
|
220
303
|
functionCall: true,
|
|
221
304
|
},
|
|
222
|
-
contextWindowTokens:
|
|
305
|
+
contextWindowTokens: 131_072,
|
|
306
|
+
description: 'Kimi K2 Instruct 是月之暗面推出的大语言模型,具有超长上下文处理能力。',
|
|
307
|
+
displayName: 'Kimi K2 Instruct',
|
|
308
|
+
id: 'kimi-k2-instruct',
|
|
309
|
+
maxOutput: 32_768,
|
|
310
|
+
pricing: {
|
|
311
|
+
currency: 'CNY',
|
|
312
|
+
units: [
|
|
313
|
+
{ name: 'textInput', rate: 4, strategy: 'fixed', unit: 'millionTokens' },
|
|
314
|
+
{ name: 'textOutput', rate: 16, strategy: 'fixed', unit: 'millionTokens' },
|
|
315
|
+
],
|
|
316
|
+
},
|
|
317
|
+
type: 'chat',
|
|
318
|
+
},
|
|
319
|
+
{
|
|
320
|
+
abilities: {
|
|
321
|
+
reasoning: true,
|
|
322
|
+
},
|
|
323
|
+
contextWindowTokens: 32_000,
|
|
324
|
+
description: 'ERNIE 4.5 300B A47B 是百度文心推出的超大规模混合专家模型,具备卓越的推理能力。',
|
|
325
|
+
displayName: 'ERNIE 4.5 300B A47B',
|
|
326
|
+
id: 'ernie-4.5-300b-a47b',
|
|
327
|
+
maxOutput: 8192,
|
|
328
|
+
pricing: {
|
|
329
|
+
currency: 'CNY',
|
|
330
|
+
units: [
|
|
331
|
+
{ name: 'textInput', rate: 2.9, strategy: 'fixed', unit: 'millionTokens' },
|
|
332
|
+
{ name: 'textOutput', rate: 2.9, strategy: 'fixed', unit: 'millionTokens' },
|
|
333
|
+
],
|
|
334
|
+
},
|
|
335
|
+
type: 'chat',
|
|
336
|
+
},
|
|
337
|
+
{
|
|
338
|
+
contextWindowTokens: 120_000,
|
|
339
|
+
description: 'ERNIE 4.5 21B A3B 是百度文心推出的混合专家模型,具备强大的推理和多语言能力。',
|
|
340
|
+
displayName: 'ERNIE 4.5 21B A3B',
|
|
341
|
+
id: 'ernie-4.5-21b-a3b',
|
|
342
|
+
maxOutput: 8192,
|
|
343
|
+
pricing: {
|
|
344
|
+
currency: 'CNY',
|
|
345
|
+
units: [
|
|
346
|
+
{ name: 'textInput', rate: 1, strategy: 'fixed', unit: 'millionTokens' },
|
|
347
|
+
{ name: 'textOutput', rate: 1, strategy: 'fixed', unit: 'millionTokens' },
|
|
348
|
+
],
|
|
349
|
+
},
|
|
350
|
+
type: 'chat',
|
|
351
|
+
},
|
|
352
|
+
{
|
|
353
|
+
abilities: {
|
|
354
|
+
functionCall: true,
|
|
355
|
+
reasoning: true,
|
|
356
|
+
},
|
|
357
|
+
contextWindowTokens: 131_072,
|
|
223
358
|
description:
|
|
224
|
-
'
|
|
225
|
-
displayName: '
|
|
226
|
-
|
|
227
|
-
|
|
359
|
+
'Qwen3-8B 是 Qwen 系列第三代的大型语言模型,拥有 82 亿参数,专为高效推理和多语言任务设计。支持无缝切换思维模式(复杂推理)和非思维模式(通用对话),在数学、编码、常识推理及多语言指令执行中表现出色。',
|
|
360
|
+
displayName: 'Qwen3 8B',
|
|
361
|
+
id: 'qwen3-8b',
|
|
362
|
+
maxOutput: 8192,
|
|
228
363
|
pricing: {
|
|
229
364
|
currency: 'CNY',
|
|
230
365
|
units: [
|
|
231
|
-
{ name: 'textInput', rate:
|
|
232
|
-
{ name: 'textOutput', rate:
|
|
366
|
+
{ name: 'textInput', rate: 1, strategy: 'fixed', unit: 'millionTokens' },
|
|
367
|
+
{ name: 'textOutput', rate: 4, strategy: 'fixed', unit: 'millionTokens' },
|
|
233
368
|
],
|
|
234
369
|
},
|
|
235
370
|
type: 'chat',
|
|
236
371
|
},
|
|
237
372
|
{
|
|
238
|
-
|
|
373
|
+
abilities: {
|
|
374
|
+
functionCall: true,
|
|
375
|
+
reasoning: true,
|
|
376
|
+
},
|
|
377
|
+
contextWindowTokens: 131_072,
|
|
239
378
|
description:
|
|
240
|
-
'
|
|
241
|
-
displayName: '
|
|
242
|
-
id: '
|
|
379
|
+
'Qwen3-14B 是 Qwen 系列第三代的大型语言模型,拥有 148 亿参数,专为高效推理和多语言任务设计。支持无缝切换思维模式(复杂推理)和非思维模式(通用对话),在数学、编码、常识推理及多语言指令执行中表现出色。',
|
|
380
|
+
displayName: 'Qwen3 14B',
|
|
381
|
+
id: 'qwen3-14b',
|
|
382
|
+
maxOutput: 8192,
|
|
243
383
|
pricing: {
|
|
244
384
|
currency: 'CNY',
|
|
245
385
|
units: [
|
|
246
|
-
{ name: 'textInput', rate:
|
|
247
|
-
{ name: 'textOutput', rate:
|
|
386
|
+
{ name: 'textInput', rate: 2, strategy: 'fixed', unit: 'millionTokens' },
|
|
387
|
+
{ name: 'textOutput', rate: 8, strategy: 'fixed', unit: 'millionTokens' },
|
|
248
388
|
],
|
|
249
389
|
},
|
|
250
390
|
type: 'chat',
|
|
251
391
|
},
|
|
252
392
|
{
|
|
253
|
-
|
|
393
|
+
abilities: {
|
|
394
|
+
functionCall: true,
|
|
395
|
+
reasoning: true,
|
|
396
|
+
},
|
|
397
|
+
contextWindowTokens: 131_072,
|
|
254
398
|
description:
|
|
255
|
-
'
|
|
256
|
-
displayName: '
|
|
257
|
-
id: '
|
|
399
|
+
'Qwen3-32B 是 Qwen 系列第三代的大型语言模型,拥有 328 亿参数,专为高效推理和多语言任务设计。支持无缝切换思考模式(复杂推理)和非思考模式(通用对话),在数学、编码、常识推理及多语言指令执行中表现出色。',
|
|
400
|
+
displayName: 'Qwen3 32B',
|
|
401
|
+
id: 'qwen3-32b',
|
|
402
|
+
maxOutput: 8192,
|
|
258
403
|
pricing: {
|
|
259
404
|
currency: 'CNY',
|
|
260
405
|
units: [
|
|
261
|
-
{ name: 'textInput', rate:
|
|
262
|
-
{ name: 'textOutput', rate:
|
|
406
|
+
{ name: 'textInput', rate: 2.9, strategy: 'fixed', unit: 'millionTokens' },
|
|
407
|
+
{ name: 'textOutput', rate: 11.6, strategy: 'fixed', unit: 'millionTokens' },
|
|
263
408
|
],
|
|
264
409
|
},
|
|
265
410
|
type: 'chat',
|
|
@@ -267,17 +412,19 @@ const infiniaiChatModels: AIChatModelCard[] = [
|
|
|
267
412
|
{
|
|
268
413
|
abilities: {
|
|
269
414
|
functionCall: true,
|
|
415
|
+
reasoning: true,
|
|
270
416
|
},
|
|
271
|
-
contextWindowTokens:
|
|
417
|
+
contextWindowTokens: 131_072,
|
|
272
418
|
description:
|
|
273
|
-
'
|
|
274
|
-
displayName: '
|
|
275
|
-
id: '
|
|
419
|
+
'Qwen3-30B-A3B 是 Qwen 系列第三代的大型语言模型,采用混合专家(MoE)架构,总计 305 亿参数,每 token 激活 33 亿参数。支持无缝切换思维模式(复杂推理)和非思维模式(通用对话),在数学、编码、常识推理及多语言指令执行中表现出色。',
|
|
420
|
+
displayName: 'Qwen3 30B A3B',
|
|
421
|
+
id: 'qwen3-30b-a3b',
|
|
422
|
+
maxOutput: 8192,
|
|
276
423
|
pricing: {
|
|
277
424
|
currency: 'CNY',
|
|
278
425
|
units: [
|
|
279
|
-
{ name: 'textInput', rate:
|
|
280
|
-
{ name: 'textOutput', rate:
|
|
426
|
+
{ name: 'textInput', rate: 2.2, strategy: 'fixed', unit: 'millionTokens' },
|
|
427
|
+
{ name: 'textOutput', rate: 8.7, strategy: 'fixed', unit: 'millionTokens' },
|
|
281
428
|
],
|
|
282
429
|
},
|
|
283
430
|
type: 'chat',
|
|
@@ -287,44 +434,156 @@ const infiniaiChatModels: AIChatModelCard[] = [
|
|
|
287
434
|
functionCall: true,
|
|
288
435
|
reasoning: true,
|
|
289
436
|
},
|
|
290
|
-
contextWindowTokens:
|
|
437
|
+
contextWindowTokens: 131_072,
|
|
438
|
+
description:
|
|
439
|
+
'Qwen3-235B-A22B 是 Qwen 系列第三代的大型语言模型,采用混合专家(MoE)架构,总计 2350 亿参数,每 token 激活 220 亿参数。支持无缝切换思考模式(复杂推理)和非思维模式(通用对话),在数学、编码、常识推理及多语言指令执行中表现出色。',
|
|
440
|
+
displayName: 'Qwen3 235B A22B',
|
|
441
|
+
id: 'qwen3-235b-a22b',
|
|
442
|
+
maxOutput: 8192,
|
|
443
|
+
pricing: {
|
|
444
|
+
currency: 'CNY',
|
|
445
|
+
units: [
|
|
446
|
+
{ name: 'textInput', rate: 5, strategy: 'fixed', unit: 'millionTokens' },
|
|
447
|
+
{ name: 'textOutput', rate: 20, strategy: 'fixed', unit: 'millionTokens' },
|
|
448
|
+
],
|
|
449
|
+
},
|
|
450
|
+
type: 'chat',
|
|
451
|
+
},
|
|
452
|
+
{
|
|
453
|
+
abilities: {
|
|
454
|
+
vision: true,
|
|
455
|
+
},
|
|
456
|
+
contextWindowTokens: 125_000,
|
|
457
|
+
description:
|
|
458
|
+
'Qwen2.5-VL 系列模型提升了模型的智能水平、实用性和适用性,使其在自然对话、内容创作、专业知识服务及代码开发等场景中表现更优。旗舰模型 Qwen2.5-VL-72B-Instruct 在涵盖多个领域和任务的基准测试中展现出强大的竞争力,包括大学水平的问题解答、数学、文档理解、通用问答、视频理解以及视觉代理任务等。',
|
|
459
|
+
displayName: 'Qwen2.5 VL 72B Instruct',
|
|
460
|
+
id: 'qwen2.5-vl-72b-instruct',
|
|
461
|
+
maxOutput: 4096,
|
|
462
|
+
pricing: {
|
|
463
|
+
currency: 'CNY',
|
|
464
|
+
units: [
|
|
465
|
+
{ name: 'textInput', rate: 10, strategy: 'fixed', unit: 'millionTokens' },
|
|
466
|
+
{ name: 'textOutput', rate: 10, strategy: 'fixed', unit: 'millionTokens' },
|
|
467
|
+
],
|
|
468
|
+
},
|
|
469
|
+
type: 'chat',
|
|
470
|
+
},
|
|
471
|
+
{
|
|
472
|
+
abilities: {
|
|
473
|
+
vision: true,
|
|
474
|
+
},
|
|
475
|
+
contextWindowTokens: 125_000,
|
|
291
476
|
description:
|
|
292
|
-
'
|
|
477
|
+
'Qwen2.5-VL 系列模型提升了模型的智能水平、实用性和适用性,使其在自然对话、内容创作、专业知识服务及代码开发等场景中表现更优。模型 Qwen2.5-VL-32B-Instruct 在涵盖多个领域和任务的基准测试中展现出强大的竞争力,包括大学水平的问题解答、数学、文档理解、通用问答、视频理解以及视觉代理任务等。',
|
|
478
|
+
displayName: 'Qwen2.5 VL 32B Instruct',
|
|
479
|
+
id: 'qwen2.5-vl-32b-instruct',
|
|
480
|
+
maxOutput: 4096,
|
|
481
|
+
pricing: {
|
|
482
|
+
currency: 'CNY',
|
|
483
|
+
units: [
|
|
484
|
+
{ name: 'textInput', rate: 5.8, strategy: 'fixed', unit: 'millionTokens' },
|
|
485
|
+
{ name: 'textOutput', rate: 5.8, strategy: 'fixed', unit: 'millionTokens' },
|
|
486
|
+
],
|
|
487
|
+
},
|
|
488
|
+
type: 'chat',
|
|
489
|
+
},
|
|
490
|
+
{
|
|
491
|
+
abilities: {
|
|
492
|
+
vision: true,
|
|
493
|
+
},
|
|
494
|
+
contextWindowTokens: 125_000,
|
|
495
|
+
description:
|
|
496
|
+
'Qwen2.5-VL 系列模型提升了模型的智能水平、实用性和适用性,使其在自然对话、内容创作、专业知识服务及代码开发等场景中表现更优。模型 Qwen2.5-VL-7B-Instruct 在涵盖多个领域和任务的基准测试中展现出强大的竞争力,包括大学水平的问题解答、数学、文档理解、通用问答、视频理解以及视觉代理任务等。',
|
|
497
|
+
displayName: 'Qwen2.5 VL 7B Instruct',
|
|
498
|
+
id: 'qwen2.5-vl-7b-instruct',
|
|
499
|
+
maxOutput: 4096,
|
|
500
|
+
pricing: {
|
|
501
|
+
currency: 'CNY',
|
|
502
|
+
units: [
|
|
503
|
+
{ name: 'textInput', rate: 2, strategy: 'fixed', unit: 'millionTokens' },
|
|
504
|
+
{ name: 'textOutput', rate: 2, strategy: 'fixed', unit: 'millionTokens' },
|
|
505
|
+
],
|
|
506
|
+
},
|
|
507
|
+
type: 'chat',
|
|
508
|
+
},
|
|
509
|
+
{
|
|
510
|
+
abilities: {
|
|
511
|
+
reasoning: true,
|
|
512
|
+
},
|
|
513
|
+
contextWindowTokens: 32_000,
|
|
514
|
+
description: 'QwQ 32B 是通义千问推出的推理专用模型,专注于推理任务。',
|
|
293
515
|
displayName: 'QwQ 32B',
|
|
294
|
-
enabled: false,
|
|
295
516
|
id: 'qwq-32b',
|
|
517
|
+
maxOutput: 8192,
|
|
296
518
|
pricing: {
|
|
297
519
|
currency: 'CNY',
|
|
298
520
|
units: [
|
|
299
|
-
{ name: 'textInput', rate:
|
|
300
|
-
{ name: 'textOutput', rate:
|
|
521
|
+
{ name: 'textInput', rate: 2.9, strategy: 'fixed', unit: 'millionTokens' },
|
|
522
|
+
{ name: 'textOutput', rate: 2.9, strategy: 'fixed', unit: 'millionTokens' },
|
|
301
523
|
],
|
|
302
524
|
},
|
|
303
525
|
type: 'chat',
|
|
304
526
|
},
|
|
305
527
|
{
|
|
306
|
-
contextWindowTokens:
|
|
528
|
+
contextWindowTokens: 131_072,
|
|
529
|
+
description:
|
|
530
|
+
'DeepSeek-V3-0324 是一个强大的专家混合(MoE)语言模型,总参数量为 671B,每个 Token 激活 37B 参数。该模型采用多头潜在注意力(MLA)和 DeepSeekMoE 架构,实现了高效推理和经济训练,并在前代 DeepSeek-V3 的基础上显著提升了性能。',
|
|
531
|
+
displayName: 'DeepSeek V3 0324',
|
|
532
|
+
id: 'deepseek-v3',
|
|
533
|
+
maxOutput: 16_384,
|
|
534
|
+
pricing: {
|
|
535
|
+
currency: 'CNY',
|
|
536
|
+
units: [
|
|
537
|
+
{ name: 'textInput', rate: 2, strategy: 'fixed', unit: 'millionTokens' },
|
|
538
|
+
{ name: 'textOutput', rate: 8, strategy: 'fixed', unit: 'millionTokens' },
|
|
539
|
+
],
|
|
540
|
+
},
|
|
541
|
+
type: 'chat',
|
|
542
|
+
},
|
|
543
|
+
{
|
|
544
|
+
abilities: {
|
|
545
|
+
functionCall: true,
|
|
546
|
+
reasoning: true,
|
|
547
|
+
},
|
|
548
|
+
contextWindowTokens: 131_072,
|
|
307
549
|
description:
|
|
308
|
-
'DeepSeek-R1
|
|
550
|
+
'DeepSeek-R1 是一个专注于推理能力的大语言模型,通过创新的训练流程实现了与 OpenAI-o1 相当的数学、代码和推理任务表现。该模型采用了冷启动数据和大规模强化学习相结合的方式进行训练。',
|
|
551
|
+
displayName: 'DeepSeek R1',
|
|
552
|
+
id: 'deepseek-r1',
|
|
553
|
+
maxOutput: 32_768,
|
|
554
|
+
pricing: {
|
|
555
|
+
currency: 'CNY',
|
|
556
|
+
units: [
|
|
557
|
+
{ name: 'textInput', rate: 4, strategy: 'fixed', unit: 'millionTokens' },
|
|
558
|
+
{ name: 'textOutput', rate: 16, strategy: 'fixed', unit: 'millionTokens' },
|
|
559
|
+
],
|
|
560
|
+
},
|
|
561
|
+
type: 'chat',
|
|
562
|
+
},
|
|
563
|
+
{
|
|
564
|
+
abilities: {
|
|
565
|
+
reasoning: true,
|
|
566
|
+
},
|
|
567
|
+
contextWindowTokens: 32_000,
|
|
568
|
+
description: 'DeepSeek R1 Distill Qwen 32B 是深度求索基于Qwen蒸馏的高效模型。',
|
|
309
569
|
displayName: 'DeepSeek R1 Distill Qwen 32B',
|
|
310
|
-
enabled: false,
|
|
311
570
|
id: 'deepseek-r1-distill-qwen-32b',
|
|
571
|
+
maxOutput: 8192,
|
|
312
572
|
pricing: {
|
|
313
573
|
currency: 'CNY',
|
|
314
574
|
units: [
|
|
315
|
-
{ name: 'textInput', rate:
|
|
316
|
-
{ name: 'textOutput', rate:
|
|
575
|
+
{ name: 'textInput', rate: 2.9, strategy: 'fixed', unit: 'millionTokens' },
|
|
576
|
+
{ name: 'textOutput', rate: 2.9, strategy: 'fixed', unit: 'millionTokens' },
|
|
317
577
|
],
|
|
318
578
|
},
|
|
319
579
|
type: 'chat',
|
|
320
580
|
},
|
|
321
581
|
{
|
|
322
|
-
contextWindowTokens:
|
|
323
|
-
description:
|
|
324
|
-
|
|
325
|
-
|
|
326
|
-
|
|
327
|
-
id: 'llama-3.3-70b-instruct',
|
|
582
|
+
contextWindowTokens: 32_000,
|
|
583
|
+
description: 'Megrez 3B Instruct 是无问芯穹推出的小参数量高效模型。',
|
|
584
|
+
displayName: 'Megrez 3B Instruct',
|
|
585
|
+
id: 'megrez-3b-instruct',
|
|
586
|
+
maxOutput: 4096,
|
|
328
587
|
pricing: {
|
|
329
588
|
currency: 'CNY',
|
|
330
589
|
units: [
|
|
@@ -335,120 +594,152 @@ const infiniaiChatModels: AIChatModelCard[] = [
|
|
|
335
594
|
type: 'chat',
|
|
336
595
|
},
|
|
337
596
|
{
|
|
597
|
+
abilities: {
|
|
598
|
+
functionCall: true,
|
|
599
|
+
},
|
|
338
600
|
contextWindowTokens: 32_768,
|
|
339
601
|
description:
|
|
340
|
-
'Qwen2 是 Qwen
|
|
341
|
-
displayName: '
|
|
342
|
-
id: 'qwen2-
|
|
602
|
+
'Qwen2.5 是 Qwen 大型语言模型系列的最新成果。Qwen2.5 发布了从 0.5 到 720 亿参数不等的基础语言模型及指令调优语言模型。Qwen2.5 相比 Qwen2 带来了以下改进:\n显著增加知识量,在编程与数学领域的能力得到极大提升。\n在遵循指令、生成长文本、理解结构化数据 (例如,表格) 以及生成结构化输出特别是 JSON 方面有显著提升。对系统提示的多样性更具韧性,增强了聊天机器人中的角色扮演实现和条件设定。\n支持长上下文处理。\n支持超过 29 种语言的多语言功能,包括中文、英语、法语、西班牙语、葡萄牙语、德语、意大利语、俄语、日语、韩语、越南语、泰语、阿拉伯语等。',
|
|
603
|
+
displayName: 'Qwen2.5 32B Instruct',
|
|
604
|
+
id: 'qwen2.5-32b-instruct',
|
|
343
605
|
pricing: {
|
|
344
606
|
currency: 'CNY',
|
|
345
607
|
units: [
|
|
346
|
-
{ name: 'textInput', rate:
|
|
347
|
-
{ name: 'textOutput', rate:
|
|
608
|
+
{ name: 'textInput', rate: 2.9, strategy: 'fixed', unit: 'millionTokens' },
|
|
609
|
+
{ name: 'textOutput', rate: 2.9, strategy: 'fixed', unit: 'millionTokens' },
|
|
348
610
|
],
|
|
349
611
|
},
|
|
350
612
|
type: 'chat',
|
|
351
613
|
},
|
|
352
614
|
{
|
|
615
|
+
abilities: {
|
|
616
|
+
functionCall: true,
|
|
617
|
+
},
|
|
353
618
|
contextWindowTokens: 32_768,
|
|
354
619
|
description:
|
|
355
|
-
'Qwen2 是 Qwen
|
|
356
|
-
displayName: '
|
|
357
|
-
id: 'qwen2-
|
|
620
|
+
'Qwen2.5 是 Qwen 大型语言模型系列的最新成果。Qwen2.5 发布了从 0.5 到 720 亿参数不等的基础语言模型及指令调优语言模型。Qwen2.5 相比 Qwen2 带来了以下改进:\n显著增加知识量,在编程与数学领域的能力得到极大提升。\n在遵循指令、生成长文本、理解结构化数据 (例如,表格) 以及生成结构化输出特别是 JSON 方面有显著提升。对系统提示的多样性更具韧性,增强了聊天机器人中的角色扮演实现和条件设定。\n支持长上下文处理。\n支持超过 29 种语言的多语言功能,包括中文、英语、法语、西班牙语、葡萄牙语、德语、意大利语、俄语、日语、韩语、越南语、泰语、阿拉伯语等。',
|
|
621
|
+
displayName: 'Qwen2.5 72B Instruct',
|
|
622
|
+
id: 'qwen2.5-72b-instruct',
|
|
358
623
|
pricing: {
|
|
359
624
|
currency: 'CNY',
|
|
360
625
|
units: [
|
|
361
|
-
{ name: 'textInput', rate:
|
|
362
|
-
{ name: 'textOutput', rate:
|
|
626
|
+
{ name: 'textInput', rate: 5, strategy: 'fixed', unit: 'millionTokens' },
|
|
627
|
+
{ name: 'textOutput', rate: 5, strategy: 'fixed', unit: 'millionTokens' },
|
|
363
628
|
],
|
|
364
629
|
},
|
|
365
630
|
type: 'chat',
|
|
366
631
|
},
|
|
367
632
|
{
|
|
368
|
-
contextWindowTokens:
|
|
633
|
+
contextWindowTokens: 32_768,
|
|
369
634
|
description:
|
|
370
|
-
'
|
|
371
|
-
displayName: '
|
|
372
|
-
id: '
|
|
635
|
+
'Qwen2.5 是 Qwen 大型语言模型系列的最新成果。Qwen2.5 发布了从 0.5 到 720 亿参数不等的基础语言模型及指令调优语言模型。Qwen2.5 相比 Qwen2 带来了以下改进:\n显著增加知识量,在编程与数学领域的能力得到极大提升。\n在遵循指令、生成长文本、理解结构化数据 (例如,表格) 以及生成结构化输出特别是 JSON 方面有显著提升。对系统提示的多样性更具韧性,增强了聊天机器人中的角色扮演实现和条件设定。\n支持长上下文处理。\n支持超过 29 种语言的多语言功能,包括中文、英语、法语、西班牙语、葡萄牙语、德语、意大利语、俄语、日语、韩语、越南语、泰语、阿拉伯语等。',
|
|
636
|
+
displayName: 'Qwen2.5 14B Instruct',
|
|
637
|
+
id: 'qwen2.5-14b-instruct',
|
|
373
638
|
pricing: {
|
|
374
639
|
currency: 'CNY',
|
|
375
640
|
units: [
|
|
376
|
-
{ name: 'textInput', rate:
|
|
377
|
-
{ name: 'textOutput', rate:
|
|
641
|
+
{ name: 'textInput', rate: 2, strategy: 'fixed', unit: 'millionTokens' },
|
|
642
|
+
{ name: 'textOutput', rate: 2, strategy: 'fixed', unit: 'millionTokens' },
|
|
378
643
|
],
|
|
379
644
|
},
|
|
380
645
|
type: 'chat',
|
|
381
646
|
},
|
|
382
647
|
{
|
|
383
|
-
|
|
384
|
-
|
|
385
|
-
|
|
386
|
-
|
|
648
|
+
abilities: {
|
|
649
|
+
functionCall: true,
|
|
650
|
+
},
|
|
651
|
+
contextWindowTokens: 32_768,
|
|
652
|
+
description:
|
|
653
|
+
'Qwen2.5 是 Qwen 大型语言模型系列的最新成果。Qwen2.5 发布了从 0.5 到 720 亿参数不等的基础语言模型及指令调优语言模型。Qwen2.5 相比 Qwen2 带来了以下改进:\n显著增加知识量,在编程与数学领域的能力得到极大提升。\n在遵循指令、生成长文本、理解结构化数据 (例如,表格) 以及生成结构化输出特别是 JSON 方面有显著提升。对系统提示的多样性更具韧性,增强了聊天机器人中的角色扮演实现和条件设定。\n支持长上下文处理。\n支持超过 29 种语言的多语言功能,包括中文、英语、法语、西班牙语、葡萄牙语、德语、意大利语、俄语、日语、韩语、越南语、泰语、阿拉伯语等。',
|
|
654
|
+
displayName: 'Qwen2.5 7B Instruct',
|
|
655
|
+
id: 'qwen2.5-7b-instruct',
|
|
387
656
|
pricing: {
|
|
388
657
|
currency: 'CNY',
|
|
389
658
|
units: [
|
|
390
|
-
{ name: 'textInput', rate:
|
|
391
|
-
{ name: 'textOutput', rate:
|
|
659
|
+
{ name: 'textInput', rate: 1, strategy: 'fixed', unit: 'millionTokens' },
|
|
660
|
+
{ name: 'textOutput', rate: 1, strategy: 'fixed', unit: 'millionTokens' },
|
|
392
661
|
],
|
|
393
662
|
},
|
|
394
663
|
type: 'chat',
|
|
395
664
|
},
|
|
396
665
|
{
|
|
397
|
-
|
|
666
|
+
abilities: {
|
|
667
|
+
functionCall: true,
|
|
668
|
+
},
|
|
669
|
+
contextWindowTokens: 32_768,
|
|
398
670
|
description:
|
|
399
|
-
'
|
|
400
|
-
displayName: '
|
|
401
|
-
id: '
|
|
671
|
+
'Qwen2.5-Coder 是最新的代码专用 Qwen 大型语言模型系列。Qwen2.5-Coder 在 CodeQwen1.5 的基础上带来了以下改进:\n显著提升代码生成、代码推理和代码修复能力。\n支持真实世界应用,例如代码代理,增强编码能力和数学及一般能力。\n支持长上下文处理。',
|
|
672
|
+
displayName: 'Qwen2.5 Coder 32B Instruct',
|
|
673
|
+
id: 'qwen2.5-coder-32b-instruct',
|
|
674
|
+
maxOutput: 4096,
|
|
402
675
|
pricing: {
|
|
403
676
|
currency: 'CNY',
|
|
404
677
|
units: [
|
|
405
|
-
{ name: 'textInput', rate:
|
|
406
|
-
{ name: 'textOutput', rate:
|
|
678
|
+
{ name: 'textInput', rate: 2.9, strategy: 'fixed', unit: 'millionTokens' },
|
|
679
|
+
{ name: 'textOutput', rate: 2.9, strategy: 'fixed', unit: 'millionTokens' },
|
|
407
680
|
],
|
|
408
681
|
},
|
|
409
682
|
type: 'chat',
|
|
410
683
|
},
|
|
411
684
|
{
|
|
412
|
-
contextWindowTokens:
|
|
413
|
-
description:
|
|
414
|
-
|
|
415
|
-
|
|
416
|
-
|
|
685
|
+
contextWindowTokens: 131_072,
|
|
686
|
+
description: '该模型需要申请体验。GPT-OSS-120B 是 OpenAI 推出的开源大规模语言模型,具备强大的文本生成能力。',
|
|
687
|
+
displayName: 'GPT-OSS-120B',
|
|
688
|
+
id: 'gpt-oss-120b',
|
|
689
|
+
maxOutput: 32_768,
|
|
417
690
|
pricing: {
|
|
418
691
|
currency: 'CNY',
|
|
419
692
|
units: [
|
|
420
|
-
{ name: 'textInput', rate: 0, strategy: 'fixed', unit: 'millionTokens' },
|
|
421
|
-
{ name: 'textOutput', rate:
|
|
693
|
+
{ name: 'textInput', rate: 0.8, strategy: 'fixed', unit: 'millionTokens' },
|
|
694
|
+
{ name: 'textOutput', rate: 3.2, strategy: 'fixed', unit: 'millionTokens' },
|
|
422
695
|
],
|
|
423
696
|
},
|
|
424
697
|
type: 'chat',
|
|
425
698
|
},
|
|
426
699
|
{
|
|
427
|
-
contextWindowTokens:
|
|
428
|
-
description:
|
|
429
|
-
|
|
430
|
-
|
|
431
|
-
|
|
700
|
+
contextWindowTokens: 131_072,
|
|
701
|
+
description: '该模型需要申请体验。GPT-OSS-20B 是 OpenAI 推出的开源中型语言模型,具备高效的文本生成能力。',
|
|
702
|
+
displayName: 'GPT-OSS-20B',
|
|
703
|
+
id: 'gpt-oss-20b',
|
|
704
|
+
maxOutput: 32_768,
|
|
432
705
|
pricing: {
|
|
433
706
|
currency: 'CNY',
|
|
434
707
|
units: [
|
|
435
|
-
{ name: 'textInput', rate: 0, strategy: 'fixed', unit: 'millionTokens' },
|
|
436
|
-
{ name: 'textOutput', rate:
|
|
708
|
+
{ name: 'textInput', rate: 0.4, strategy: 'fixed', unit: 'millionTokens' },
|
|
709
|
+
{ name: 'textOutput', rate: 1.6, strategy: 'fixed', unit: 'millionTokens' },
|
|
437
710
|
],
|
|
438
711
|
},
|
|
439
712
|
type: 'chat',
|
|
440
713
|
},
|
|
441
714
|
{
|
|
442
|
-
|
|
443
|
-
|
|
444
|
-
|
|
445
|
-
|
|
446
|
-
|
|
715
|
+
abilities: {
|
|
716
|
+
reasoning: true,
|
|
717
|
+
},
|
|
718
|
+
contextWindowTokens: 131_072,
|
|
719
|
+
description: '企业专属服务专用模型,包并发服务。',
|
|
720
|
+
displayName: 'DeepSeek R1 (Pro)',
|
|
721
|
+
id: 'pro-deepseek-r1',
|
|
722
|
+
maxOutput: 32_768,
|
|
447
723
|
pricing: {
|
|
448
724
|
currency: 'CNY',
|
|
449
725
|
units: [
|
|
450
|
-
{ name: 'textInput', rate:
|
|
451
|
-
{ name: 'textOutput', rate:
|
|
726
|
+
{ name: 'textInput', rate: 4, strategy: 'fixed', unit: 'millionTokens' },
|
|
727
|
+
{ name: 'textOutput', rate: 16, strategy: 'fixed', unit: 'millionTokens' },
|
|
728
|
+
],
|
|
729
|
+
},
|
|
730
|
+
type: 'chat',
|
|
731
|
+
},
|
|
732
|
+
{
|
|
733
|
+
contextWindowTokens: 131_072,
|
|
734
|
+
description: '企业专属服务专用模型,包并发服务。',
|
|
735
|
+
displayName: 'DeepSeek V3 (Pro)',
|
|
736
|
+
id: 'pro-deepseek-v3',
|
|
737
|
+
maxOutput: 16_384,
|
|
738
|
+
pricing: {
|
|
739
|
+
currency: 'CNY',
|
|
740
|
+
units: [
|
|
741
|
+
{ name: 'textInput', rate: 2, strategy: 'fixed', unit: 'millionTokens' },
|
|
742
|
+
{ name: 'textOutput', rate: 8, strategy: 'fixed', unit: 'millionTokens' },
|
|
452
743
|
],
|
|
453
744
|
},
|
|
454
745
|
type: 'chat',
|
|
@@ -26,10 +26,10 @@ const modelscopeChatModels: AIChatModelCard[] = [
|
|
|
26
26
|
reasoning: true,
|
|
27
27
|
},
|
|
28
28
|
contextWindowTokens: 131_072,
|
|
29
|
-
description: 'DeepSeek V3.
|
|
30
|
-
displayName: 'DeepSeek V3.
|
|
29
|
+
description: 'DeepSeek V3.2 Exp 模型为混合推理架构模型,同时支持思考模式与非思考模式。',
|
|
30
|
+
displayName: 'DeepSeek V3.2 Exp',
|
|
31
31
|
enabled: true,
|
|
32
|
-
id: 'deepseek-ai/DeepSeek-V3.
|
|
32
|
+
id: 'deepseek-ai/DeepSeek-V3.2-Exp',
|
|
33
33
|
type: 'chat',
|
|
34
34
|
},
|
|
35
35
|
{
|
|
@@ -38,10 +38,9 @@ const modelscopeChatModels: AIChatModelCard[] = [
|
|
|
38
38
|
reasoning: true,
|
|
39
39
|
},
|
|
40
40
|
contextWindowTokens: 131_072,
|
|
41
|
-
description:
|
|
42
|
-
|
|
43
|
-
|
|
44
|
-
id: 'deepseek-ai/DeepSeek-R1-0528',
|
|
41
|
+
description: 'DeepSeek V3.1 模型为混合推理架构模型,同时支持思考模式与非思考模式。',
|
|
42
|
+
displayName: 'DeepSeek V3.1',
|
|
43
|
+
id: 'deepseek-ai/DeepSeek-V3.1',
|
|
45
44
|
type: 'chat',
|
|
46
45
|
},
|
|
47
46
|
{
|
|
@@ -50,19 +49,10 @@ const modelscopeChatModels: AIChatModelCard[] = [
|
|
|
50
49
|
reasoning: true,
|
|
51
50
|
},
|
|
52
51
|
contextWindowTokens: 131_072,
|
|
53
|
-
description:
|
|
54
|
-
|
|
55
|
-
|
|
56
|
-
|
|
57
|
-
},
|
|
58
|
-
{
|
|
59
|
-
abilities: {
|
|
60
|
-
functionCall: true,
|
|
61
|
-
},
|
|
62
|
-
contextWindowTokens: 131_072,
|
|
63
|
-
description: 'DeepSeek V3是DeepSeek第三代模型的最新版本,具有强大的推理和对话能力。',
|
|
64
|
-
displayName: 'DeepSeek V3',
|
|
65
|
-
id: 'deepseek-ai/DeepSeek-V3',
|
|
52
|
+
description:
|
|
53
|
+
'DeepSeek R1 通过利用增加的计算资源和在后训练过程中引入算法优化机制,显著提高了其推理和推断能力的深度。该模型在各种基准评估中表现出色,包括数学、编程和一般逻辑方面。其整体性能现已接近领先模型,如 O3 和 Gemini 2.5 Pro。',
|
|
54
|
+
displayName: 'DeepSeek R1 0528',
|
|
55
|
+
id: 'deepseek-ai/DeepSeek-R1-0528',
|
|
66
56
|
type: 'chat',
|
|
67
57
|
},
|
|
68
58
|
{
|
|
@@ -191,8 +191,8 @@ const novitaChatModels: AIChatModelCard[] = [
|
|
|
191
191
|
id: 'openai/gpt-oss-20b',
|
|
192
192
|
pricing: {
|
|
193
193
|
units: [
|
|
194
|
-
{ name: 'textInput', rate: 0.
|
|
195
|
-
{ name: 'textOutput', rate: 0.
|
|
194
|
+
{ name: 'textInput', rate: 0.04, strategy: 'fixed', unit: 'millionTokens' },
|
|
195
|
+
{ name: 'textOutput', rate: 0.15, strategy: 'fixed', unit: 'millionTokens' },
|
|
196
196
|
],
|
|
197
197
|
},
|
|
198
198
|
type: 'chat',
|
|
@@ -1,9 +1,7 @@
|
|
|
1
|
-
import type { ChatModelCard } from '@lobechat/types';
|
|
2
1
|
import { ModelProvider } from 'model-bank';
|
|
3
2
|
|
|
4
3
|
import { createOpenAICompatibleRuntime } from '../../core/openaiCompatibleFactory';
|
|
5
|
-
import {
|
|
6
|
-
import { AgentRuntimeErrorType } from '../../types/error';
|
|
4
|
+
import { processMultiProviderModelList } from '../../utils/modelParse';
|
|
7
5
|
|
|
8
6
|
export interface InfiniAIModelCard {
|
|
9
7
|
id: string;
|
|
@@ -12,68 +10,24 @@ export interface InfiniAIModelCard {
|
|
|
12
10
|
export const LobeInfiniAI = createOpenAICompatibleRuntime({
|
|
13
11
|
baseURL: 'https://cloud.infini-ai.com/maas/v1',
|
|
14
12
|
chatCompletion: {
|
|
15
|
-
|
|
16
|
-
|
|
17
|
-
if (error instanceof Response) {
|
|
18
|
-
errorResponse = error;
|
|
19
|
-
} else if ('status' in (error as any)) {
|
|
20
|
-
errorResponse = error as Response;
|
|
21
|
-
}
|
|
22
|
-
if (errorResponse) {
|
|
23
|
-
if (errorResponse.status === 401) {
|
|
24
|
-
return {
|
|
25
|
-
error,
|
|
26
|
-
errorType: AgentRuntimeErrorType.InvalidProviderAPIKey,
|
|
27
|
-
};
|
|
28
|
-
}
|
|
13
|
+
handlePayload: (payload) => {
|
|
14
|
+
const { model, thinking, ...rest } = payload;
|
|
29
15
|
|
|
30
|
-
if (errorResponse.status === 429) {
|
|
31
|
-
return {
|
|
32
|
-
error,
|
|
33
|
-
errorType: AgentRuntimeErrorType.QuotaLimitReached,
|
|
34
|
-
};
|
|
35
|
-
}
|
|
36
|
-
}
|
|
37
16
|
return {
|
|
38
|
-
|
|
39
|
-
|
|
17
|
+
...rest,
|
|
18
|
+
enable_thinking: thinking !== undefined ? thinking.type === 'enabled' : false,
|
|
19
|
+
model,
|
|
20
|
+
} as any;
|
|
40
21
|
},
|
|
41
22
|
},
|
|
42
23
|
debug: {
|
|
43
24
|
chatCompletion: () => process.env.DEBUG_INFINIAI_CHAT_COMPLETION === '1',
|
|
44
25
|
},
|
|
45
26
|
models: async ({ client }) => {
|
|
46
|
-
const { LOBE_DEFAULT_MODEL_LIST } = await import('model-bank');
|
|
47
|
-
|
|
48
|
-
const reasoningKeywords = ['deepseek-r1', 'qwq', 'qwen3'];
|
|
49
|
-
const visionKeywords = ['qwen2.5-vl'];
|
|
50
|
-
|
|
51
27
|
const modelsPage = (await client.models.list()) as any;
|
|
52
28
|
const modelList: InfiniAIModelCard[] = modelsPage.data;
|
|
53
29
|
|
|
54
|
-
return modelList
|
|
55
|
-
.map((model) => {
|
|
56
|
-
const knownModel = LOBE_DEFAULT_MODEL_LIST.find(
|
|
57
|
-
(m) => model.id.toLowerCase() === m.id.toLowerCase(),
|
|
58
|
-
);
|
|
59
|
-
|
|
60
|
-
return {
|
|
61
|
-
contextWindowTokens: knownModel?.contextWindowTokens ?? undefined,
|
|
62
|
-
displayName: knownModel?.displayName ?? undefined,
|
|
63
|
-
enabled: knownModel?.enabled || false,
|
|
64
|
-
functionCall: knownModel?.abilities?.functionCall || false,
|
|
65
|
-
id: model.id,
|
|
66
|
-
reasoning:
|
|
67
|
-
reasoningKeywords.some((keyword) => model.id.toLowerCase().includes(keyword)) ||
|
|
68
|
-
knownModel?.abilities?.reasoning ||
|
|
69
|
-
false,
|
|
70
|
-
vision:
|
|
71
|
-
visionKeywords.some((keyword) => model.id.toLowerCase().includes(keyword)) ||
|
|
72
|
-
knownModel?.abilities?.vision ||
|
|
73
|
-
false,
|
|
74
|
-
};
|
|
75
|
-
})
|
|
76
|
-
.filter(Boolean) as ChatModelCard[];
|
|
30
|
+
return processMultiProviderModelList(modelList, 'infiniai');
|
|
77
31
|
},
|
|
78
32
|
provider: ModelProvider.InfiniAI,
|
|
79
33
|
});
|
|
@@ -15,6 +15,40 @@ import ProviderDetail from '../default';
|
|
|
15
15
|
|
|
16
16
|
const providerKey: GlobalLLMProviderKey = 'bedrock';
|
|
17
17
|
|
|
18
|
+
const AWS_REGIONS: string[] = [
|
|
19
|
+
'us-east-1',
|
|
20
|
+
'us-east-2',
|
|
21
|
+
'us-west-1',
|
|
22
|
+
'us-west-2',
|
|
23
|
+
'ca-central-1',
|
|
24
|
+
'us-gov-east-1',
|
|
25
|
+
'us-gov-west-1',
|
|
26
|
+
'sa-east-1',
|
|
27
|
+
'eu-north-1',
|
|
28
|
+
'eu-west-1',
|
|
29
|
+
'eu-west-2',
|
|
30
|
+
'eu-west-3',
|
|
31
|
+
'eu-central-1',
|
|
32
|
+
'eu-central-2',
|
|
33
|
+
'eu-south-1',
|
|
34
|
+
'eu-south-2',
|
|
35
|
+
'me-south-1',
|
|
36
|
+
'me-central-1',
|
|
37
|
+
'af-south-1',
|
|
38
|
+
'ap-south-1',
|
|
39
|
+
'ap-south-2',
|
|
40
|
+
'ap-east-1',
|
|
41
|
+
'ap-southeast-1',
|
|
42
|
+
'ap-southeast-2',
|
|
43
|
+
'ap-southeast-3',
|
|
44
|
+
'ap-southeast-4',
|
|
45
|
+
'ap-northeast-1',
|
|
46
|
+
'ap-northeast-2',
|
|
47
|
+
'ap-northeast-3',
|
|
48
|
+
'cn-north-1',
|
|
49
|
+
'cn-northwest-1',
|
|
50
|
+
];
|
|
51
|
+
|
|
18
52
|
const useBedrockCard = (): ProviderItem => {
|
|
19
53
|
const { t } = useTranslation('modelProvider');
|
|
20
54
|
|
|
@@ -68,11 +102,11 @@ const useBedrockCard = (): ProviderItem => {
|
|
|
68
102
|
) : (
|
|
69
103
|
<Select
|
|
70
104
|
allowClear
|
|
71
|
-
options={
|
|
105
|
+
options={AWS_REGIONS.map((i) => ({
|
|
72
106
|
label: i,
|
|
73
107
|
value: i,
|
|
74
108
|
}))}
|
|
75
|
-
placeholder={
|
|
109
|
+
placeholder={AWS_REGIONS[0]}
|
|
76
110
|
/>
|
|
77
111
|
),
|
|
78
112
|
desc: t(`${providerKey}.region.desc`),
|
|
@@ -76,32 +76,8 @@ const InfiniAI: ModelProviderCard = {
|
|
|
76
76
|
enabled: true,
|
|
77
77
|
id: 'qwen2.5-7b-instruct',
|
|
78
78
|
},
|
|
79
|
-
{
|
|
80
|
-
contextWindowTokens: 32_768,
|
|
81
|
-
description:
|
|
82
|
-
'Qwen2 是 Qwen 团队推出的新一代大型语言模型系列。它基于 Transformer 架构,并采用 SwiGLU 激活函数、注意力 QKV 偏置(attention QKV bias)、群组查询注意力(group query attention)、滑动窗口注意力(mixture of sliding window attention)与全注意力的混合等技术。此外,Qwen 团队还改进了适应多种自然语言和代码的分词器。',
|
|
83
|
-
displayName: 'Qwen 2 72B Instruct',
|
|
84
|
-
enabled: true,
|
|
85
|
-
id: 'qwen2-72b-instruct',
|
|
86
|
-
},
|
|
87
|
-
{
|
|
88
|
-
contextWindowTokens: 32_768,
|
|
89
|
-
description:
|
|
90
|
-
'Qwen2 是 Qwen 团队推出的新一代大型语言模型系列。它基于 Transformer 架构,并采用 SwiGLU 激活函数、注意力 QKV 偏置(attention QKV bias)、群组查询注意力(group query attention)、滑动窗口注意力(mixture of sliding window attention)与全注意力的混合等技术。此外,Qwen 团队还改进了适应多种自然语言和代码的分词器。',
|
|
91
|
-
displayName: 'Qwen 2 7B Instruct',
|
|
92
|
-
enabled: true,
|
|
93
|
-
id: 'qwen2-7b-instruct',
|
|
94
|
-
},
|
|
95
|
-
{
|
|
96
|
-
contextWindowTokens: 4096,
|
|
97
|
-
description:
|
|
98
|
-
'Yi-1.5 是 Yi 的升级版本。 它使用 500B Tokens 的高质量语料库在 Yi 上持续进行预训练,并在 3M 个多样化的微调样本上进行微调。',
|
|
99
|
-
displayName: 'Yi-1.5 34B Chat',
|
|
100
|
-
enabled: true,
|
|
101
|
-
id: 'yi-1.5-34b-chat',
|
|
102
|
-
},
|
|
103
79
|
],
|
|
104
|
-
checkModel: '
|
|
80
|
+
checkModel: 'qwen3-8b',
|
|
105
81
|
description:
|
|
106
82
|
'为应用开发者提供高性能、易上手、安全可靠的大模型服务,覆盖从大模型开发到大模型服务化部署的全流程。',
|
|
107
83
|
id: 'infiniai',
|
|
@@ -109,6 +85,7 @@ const InfiniAI: ModelProviderCard = {
|
|
|
109
85
|
modelsUrl: 'https://cloud.infini-ai.com/genstudio/model',
|
|
110
86
|
name: 'InfiniAI',
|
|
111
87
|
settings: {
|
|
88
|
+
disableBrowserRequest: true,
|
|
112
89
|
proxyUrl: {
|
|
113
90
|
placeholder: 'https://cloud.infini-ai.com/maas/v1',
|
|
114
91
|
},
|
|
@@ -12,22 +12,6 @@ const ModelScope: ModelProviderCard = {
|
|
|
12
12
|
functionCall: true,
|
|
13
13
|
id: 'deepseek-ai/DeepSeek-R1-0528',
|
|
14
14
|
},
|
|
15
|
-
{
|
|
16
|
-
contextWindowTokens: 131_072,
|
|
17
|
-
description: 'DeepSeek-V3是DeepSeek第三代模型的最新版本,具有强大的推理和对话能力。',
|
|
18
|
-
displayName: 'DeepSeek-V3',
|
|
19
|
-
enabled: true,
|
|
20
|
-
functionCall: true,
|
|
21
|
-
id: 'deepseek-ai/DeepSeek-V3',
|
|
22
|
-
},
|
|
23
|
-
{
|
|
24
|
-
contextWindowTokens: 131_072,
|
|
25
|
-
description: 'DeepSeek-R1是DeepSeek最新的推理模型,专注于复杂推理任务。',
|
|
26
|
-
displayName: 'DeepSeek-R1',
|
|
27
|
-
enabled: true,
|
|
28
|
-
functionCall: true,
|
|
29
|
-
id: 'deepseek-ai/DeepSeek-R1',
|
|
30
|
-
},
|
|
31
15
|
{
|
|
32
16
|
contextWindowTokens: 131_072,
|
|
33
17
|
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。',
|
|
@@ -45,7 +29,7 @@ const ModelScope: ModelProviderCard = {
|
|
|
45
29
|
id: 'Qwen/Qwen3-32B',
|
|
46
30
|
},
|
|
47
31
|
],
|
|
48
|
-
checkModel: 'Qwen/Qwen3-
|
|
32
|
+
checkModel: 'Qwen/Qwen3-4B',
|
|
49
33
|
description: 'ModelScope是阿里云推出的模型即服务平台,提供丰富的AI模型和推理服务。',
|
|
50
34
|
id: 'modelscope',
|
|
51
35
|
modelList: { showModelFetcher: true },
|