Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Commit f58c980

Browse files
sxjerusxjeru
authored
πŸ› fix: Add ExtendParamsTypeSchema for enhanced model settings (#11437)
* πŸ› fix: Update reasoning handling in OpenRouter and VercelAIGateway to include thinkingLevel and adjust gpt-5 reasoning parameters * πŸ› fix: Add ExtendParamsTypeSchema and AiModelSettingsSchema for enhanced model settings * πŸ› fix: Add ModelSearchImplementTypeSchema and update AiModelSettingsSchema for enhanced model configuration * delete gemini-2.5-flash-image-preview model * Add GLM-4.7 model to volcengine and remove deprecated GLM-4 32B 0414 model from wenxin * ✨ feat: 添加 MiniMax-M2.1 ε’Œ GLM-4.7-Flash ζ¨‘εž‹εˆ°ζ¨‘εž‹εΊ“ * ✨ feat: ζ›΄ζ–° Zhipu ζ¨‘εž‹εΊ“οΌŒζ·»εŠ  GLM-4.7-FlashX ζ¨‘εž‹εΉΆη§»ι™€ GLM-4.5-Flash ζ¨‘εž‹ * test: add extendParams mapping for gpt-5.x reasoning models in VercelAIGatewayAI * remove deprecated DeepSeek R1 model from nvidiaChatModels * i18n: ζ›΄ζ–° MiniMax-M2.1 ζ¨‘εž‹ζθΏ°δΈΊθ‹±ζ–‡
1 parent 8d00af4 commit f58c980

File tree

14 files changed

+421
-170
lines changed

14 files changed

+421
-170
lines changed

β€Žpackages/model-bank/src/aiModels/cerebras.tsβ€Ž

Lines changed: 2 additions & 22 deletions
Original file line numberDiff line numberDiff line change
@@ -1,27 +1,6 @@
11
import { AIChatModelCard } from '../types/aiModel';
22

33
const cerebrasModels: AIChatModelCard[] = [
4-
{
5-
abilities: {
6-
functionCall: true,
7-
reasoning: true,
8-
structuredOutput: true,
9-
},
10-
contextWindowTokens: 131_072,
11-
description:
12-
'Performs well on coding and reasoning tasks, supports streaming and tool calls, and fits agentic coding and complex reasoning.',
13-
displayName: 'GLM-4.6',
14-
enabled: true,
15-
id: 'zai-glm-4.6',
16-
maxOutput: 40_000,
17-
pricing: {
18-
units: [
19-
{ name: 'textInput', rate: 2.25, strategy: 'fixed', unit: 'millionTokens' },
20-
{ name: 'textOutput', rate: 2.75, strategy: 'fixed', unit: 'millionTokens' },
21-
],
22-
},
23-
type: 'chat',
24-
},
254
{
265
abilities: {
276
functionCall: true,
@@ -96,7 +75,8 @@ const cerebrasModels: AIChatModelCard[] = [
9675
functionCall: true,
9776
},
9877
contextWindowTokens: 32_768,
99-
description: 'Llama 3.1 8B: a small, low-latency Llama variant for lightweight online inference and chat.',
78+
description:
79+
'Llama 3.1 8B: a small, low-latency Llama variant for lightweight online inference and chat.',
10080
displayName: 'Llama 3.1 8B',
10181
id: 'llama3.1-8b',
10282
pricing: {

β€Žpackages/model-bank/src/aiModels/google.tsβ€Ž

Lines changed: 1 addition & 44 deletions
Original file line numberDiff line numberDiff line change
@@ -1,4 +1,4 @@
1-
import { CHAT_MODEL_IMAGE_GENERATION_PARAMS, ModelParamsSchema } from '../standard-parameters';
1+
import { ModelParamsSchema } from '../standard-parameters';
22
import { AIChatModelCard, AIImageModelCard } from '../types';
33

44
/**
@@ -485,32 +485,6 @@ const googleChatModels: AIChatModelCard[] = [
485485
},
486486
type: 'chat',
487487
},
488-
{
489-
abilities: {
490-
imageOutput: true,
491-
vision: true,
492-
},
493-
contextWindowTokens: 32_768 + 8192,
494-
description:
495-
'Nano Banana is Google’s newest, fastest, and most efficient native multimodal model, enabling conversational image generation and editing.',
496-
displayName: 'Nano Banana (Preview)',
497-
id: 'gemini-2.5-flash-image-preview',
498-
maxOutput: 8192,
499-
pricing: {
500-
approximatePricePerImage: 0.039,
501-
units: [
502-
{ name: 'textInput', rate: 0.3, strategy: 'fixed', unit: 'millionTokens' },
503-
{ name: 'imageInput', rate: 0.3, strategy: 'fixed', unit: 'millionTokens' },
504-
{ name: 'textOutput', rate: 2.5, strategy: 'fixed', unit: 'millionTokens' },
505-
{ name: 'imageOutput', rate: 30, strategy: 'fixed', unit: 'millionTokens' },
506-
],
507-
},
508-
releasedAt: '2025-08-26',
509-
settings: {
510-
extendParams: ['imageAspectRatio'],
511-
},
512-
type: 'chat',
513-
},
514488
{
515489
abilities: {
516490
functionCall: true,
@@ -967,23 +941,6 @@ const googleImageModels: AIImageModelCard[] = [
967941
],
968942
},
969943
},
970-
{
971-
displayName: 'Nano Banana (Preview)',
972-
id: 'gemini-2.5-flash-image-preview:image',
973-
type: 'image',
974-
description:
975-
'Nano Banana is Google’s newest, fastest, and most efficient native multimodal model, enabling conversational image generation and editing.',
976-
releasedAt: '2025-08-26',
977-
parameters: CHAT_MODEL_IMAGE_GENERATION_PARAMS,
978-
pricing: {
979-
approximatePricePerImage: 0.039,
980-
units: [
981-
{ name: 'textInput', rate: 0.3, strategy: 'fixed', unit: 'millionTokens' },
982-
{ name: 'textOutput', rate: 2.5, strategy: 'fixed', unit: 'millionTokens' },
983-
{ name: 'imageOutput', rate: 30, strategy: 'fixed', unit: 'millionTokens' },
984-
],
985-
},
986-
},
987944
{
988945
displayName: 'Imagen 4',
989946
id: 'imagen-4.0-generate-001',

β€Žpackages/model-bank/src/aiModels/nvidia.tsβ€Ž

Lines changed: 12 additions & 16 deletions
Original file line numberDiff line numberDiff line change
@@ -79,7 +79,8 @@ const nvidiaChatModels: AIChatModelCard[] = [
7979
vision: true,
8080
},
8181
contextWindowTokens: 128_000,
82-
description: 'A frontier vision-language model that excels at high-quality reasoning from images.',
82+
description:
83+
'A frontier vision-language model that excels at high-quality reasoning from images.',
8384
displayName: 'Llama 3.2 11B Vision Instruct',
8485
id: 'meta/llama-3.2-11b-vision-instruct',
8586
type: 'chat',
@@ -89,7 +90,8 @@ const nvidiaChatModels: AIChatModelCard[] = [
8990
vision: true,
9091
},
9192
contextWindowTokens: 128_000,
92-
description: 'A frontier vision-language model that excels at high-quality reasoning from images.',
93+
description:
94+
'A frontier vision-language model that excels at high-quality reasoning from images.',
9395
displayName: 'Llama 3.2 90B Vision Instruct',
9496
id: 'meta/llama-3.2-90b-vision-instruct',
9597
type: 'chat',
@@ -151,41 +153,35 @@ const nvidiaChatModels: AIChatModelCard[] = [
151153
},
152154
{
153155
contextWindowTokens: 8192,
154-
description: 'A frontier text generation model strong in understanding, transformation, and code generation.',
156+
description:
157+
'A frontier text generation model strong in understanding, transformation, and code generation.',
155158
displayName: 'Gemma 2 9B Instruct',
156159
id: 'google/gemma-2-9b-it',
157160
type: 'chat',
158161
},
159162
{
160163
contextWindowTokens: 8192,
161-
description: 'A frontier text generation model strong in understanding, transformation, and code generation.',
164+
description:
165+
'A frontier text generation model strong in understanding, transformation, and code generation.',
162166
displayName: 'Gemma 2 27B Instruct',
163167
id: 'google/gemma-2-27b-it',
164168
type: 'chat',
165169
},
166-
{
167-
abilities: {
168-
reasoning: true,
169-
},
170-
contextWindowTokens: 128_000,
171-
description: 'A state-of-the-art efficient LLM strong in reasoning, math, and programming.',
172-
displayName: 'DeepSeek R1',
173-
id: 'deepseek-ai/deepseek-r1',
174-
type: 'chat',
175-
},
176170
{
177171
abilities: {
178172
functionCall: true,
179173
},
180174
contextWindowTokens: 32_768,
181-
description: 'A bilingual LLM for Chinese and English across language, coding, math, and reasoning.',
175+
description:
176+
'A bilingual LLM for Chinese and English across language, coding, math, and reasoning.',
182177
displayName: 'Qwen2.5 7B Instruct',
183178
id: 'qwen/qwen2.5-7b-instruct',
184179
type: 'chat',
185180
},
186181
{
187182
contextWindowTokens: 32_768,
188-
description: 'A strong mid-sized code model with 32K context, excelling at multilingual programming.',
183+
description:
184+
'A strong mid-sized code model with 32K context, excelling at multilingual programming.',
189185
displayName: 'Qwen2.5 Coder 7B Instruct',
190186
id: 'qwen/qwen2.5-coder-7b-instruct',
191187
type: 'chat',

β€Žpackages/model-bank/src/aiModels/siliconcloud.tsβ€Ž

Lines changed: 20 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -2,6 +2,26 @@ import { AIChatModelCard, AIImageModelCard } from '../types/aiModel';
22

33
// https://siliconflow.cn/zh-cn/models
44
const siliconcloudChatModels: AIChatModelCard[] = [
5+
{
6+
abilities: {
7+
functionCall: true,
8+
reasoning: true,
9+
},
10+
contextWindowTokens: 192_000,
11+
description:
12+
'MiniMax-M2.1 is an open-source large language model optimized for agent capabilities, excelling in programming, tool usage, instruction following, and long-term planning. The model supports multilingual software development and complex multi-step workflow execution, achieving a score of 74.0 on SWE-bench Verified and surpassing Claude Sonnet 4.5 in multilingual scenarios.',
13+
displayName: 'MiniMax-M2.1 (Pro)',
14+
id: 'Pro/MiniMaxAI/MiniMax-M2.1',
15+
pricing: {
16+
currency: 'CNY',
17+
units: [
18+
{ name: 'textInput', rate: 2.1, strategy: 'fixed', unit: 'millionTokens' },
19+
{ name: 'textOutput', rate: 8.4, strategy: 'fixed', unit: 'millionTokens' },
20+
],
21+
},
22+
releasedAt: '2025-12-23',
23+
type: 'chat',
24+
},
525
{
626
abilities: {
727
functionCall: true,

β€Žpackages/model-bank/src/aiModels/volcengine.tsβ€Ž

Lines changed: 69 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -117,6 +117,75 @@ const doubaoChatModels: AIChatModelCard[] = [
117117
},
118118
type: 'chat',
119119
},
120+
{
121+
abilities: {
122+
functionCall: true,
123+
reasoning: true,
124+
},
125+
config: {
126+
deploymentName: 'glm-4-7-251222',
127+
},
128+
contextWindowTokens: 200_000,
129+
description:
130+
'GLM-4.7 is the latest flagship model from Zhipu AI. GLM-4.7 enhances coding capabilities, long-term task planning, and tool collaboration for Agentic Coding scenarios, achieving leading performance among open-source models in multiple public benchmarks. General capabilities are improved, with more concise and natural responses, and more immersive writing. In complex agent tasks, instruction following is stronger during tool calls, and the aesthetics of Artifacts and Agentic Coding frontend, as well as long-term task completion efficiency, are further enhanced. β€’ Stronger programming capabilities: Significantly improved multi-language coding and terminal agent performance; GLM-4.7 can now implement "think first, then act" mechanisms in programming frameworks like Claude Code, Kilo Code, TRAE, Cline, and Roo Code, with more stable performance on complex tasks. β€’ Frontend aesthetics improvement: GLM-4.7 shows significant progress in frontend generation quality, capable of generating websites, PPTs, and posters with better visual appeal. β€’ Stronger tool calling capabilities: GLM-4.7 enhances tool calling abilities, scoring 67 in BrowseComp web task evaluation; achieving 84.7 in τ²-Bench interactive tool calling evaluation, surpassing Claude Sonnet 4.5 as the open-source SOTA. β€’ Reasoning capability improvement: Significantly enhanced math and reasoning abilities, scoring 42.8% in the HLE ("Humanity\'s Last Exam") benchmark, a 41% improvement over GLM-4.6, surpassing GPT-5.1. β€’ General capability enhancement: GLM-4.7 conversations are more concise, intelligent, and humane; writing and role-playing are more literary and immersive.',
131+
displayName: 'GLM-4.7',
132+
id: 'glm-4-7',
133+
maxOutput: 128_000,
134+
pricing: {
135+
currency: 'CNY',
136+
units: [
137+
{
138+
lookup: {
139+
prices: {
140+
'[0, 0.032]_[0, 0.0002]': 2,
141+
'[0, 0.032]_[0.0002, infinity]': 3,
142+
'[0.032, 0.2]_[0, infinity]': 4,
143+
},
144+
pricingParams: ['textInputRange', 'textOutputRange'],
145+
},
146+
name: 'textInput',
147+
strategy: 'lookup',
148+
unit: 'millionTokens',
149+
},
150+
{
151+
lookup: {
152+
prices: {
153+
'[0, 0.032]_[0, 0.0002]': 8,
154+
'[0, 0.032]_[0.0002, infinity]': 14,
155+
'[0.032, 0.2]_[0, infinity]': 16,
156+
},
157+
pricingParams: ['textInputRange', 'textOutputRange'],
158+
},
159+
name: 'textOutput',
160+
strategy: 'lookup',
161+
unit: 'millionTokens',
162+
},
163+
{
164+
lookup: {
165+
prices: {
166+
'[0, 0.032]_[0, 0.0002]': 0.4,
167+
'[0, 0.032]_[0.0002, infinity]': 0.6,
168+
'[0.032, 0.2]_[0, infinity]': 0.8,
169+
},
170+
pricingParams: ['textInputRange', 'textOutputRange'],
171+
},
172+
name: 'textInput_cacheRead',
173+
strategy: 'lookup',
174+
unit: 'millionTokens',
175+
},
176+
{
177+
lookup: { prices: { '1h': 0.017 }, pricingParams: ['ttl'] },
178+
name: 'textInput_cacheWrite',
179+
strategy: 'lookup',
180+
unit: 'millionTokens',
181+
},
182+
],
183+
},
184+
settings: {
185+
extendParams: ['enableReasoning'],
186+
},
187+
type: 'chat',
188+
},
120189
{
121190
abilities: {
122191
functionCall: true,

0 commit comments

Comments
Β (0)