Model Provider List
International Providers
OpenAI
Provider of GPT series models, offering powerful AI capabilities. Supports GPT-5, GPT-4.1, GPT-4o and other latest models.
View Configuration Guide →
Anthropic
Provider of Claude series models, focused on safety and long context. Offers Claude Sonnet 4.5 (world's strongest coding model), Claude Haiku 4.5, Claude Opus 4.1 and other models.
View Configuration Guide →
Google Gemini
Multimodal large model launched by Google. Supports Gemini 2.0, Gemini 1.5 Pro and other models.
View Configuration Guide →
Azure OpenAI
OpenAI service on Microsoft Azure platform. Enterprise-grade deployment with strong compliance.
View Configuration Guide →
Amazon Bedrock
Large model service platform provided by AWS. Supports Claude, Llama, Mistral and many other models.
View Configuration Guide →Chinese Providers
Zhipu AI
Tsinghua-affiliated AI company offering GLM series models. Supports GLM-4, GLM-3 and other models.
View Configuration Guide →
Alibaba Cloud Bailian
Enterprise-grade large model service platform launched by Alibaba Cloud. Offers Qwen series models.
View Configuration Guide →
DeepSeek
High-performance large model launched by DeepSeek. Supports DeepSeek-R1, DeepSeek-Chat and other models.
View Configuration Guide →
Kimi
Kimi intelligent assistant launched by Moonshot AI. Supports ultra-long context understanding.
View Configuration Guide →
Tencent Hunyuan
Tencent's self-developed large language model. Supports Hunyuan-Pro, Hunyuan-Standard and other models.
View Configuration Guide →
Tencent Cloud
Large model service provided by Tencent Cloud. Supports DeepSeek-V3, Hunyuan-Pro and other models.
View Configuration Guide →
iFlytek Spark
Spark cognitive large model launched by iFlytek. Supports 4.0 Ultra, 3.5 and multiple versions.
View Configuration Guide →
Volcengine
Doubao large model service under ByteDance. Offers high-performance AI inference capabilities.
View Configuration Guide →
SiliconFlow
Service platform focused on AI inference acceleration. Supports multiple open source models.
View Configuration Guide →
Baidu Qianfan
Large language model platform launched by Baidu. Supports ERNIE-4.0, ERNIE-3.5 and other models.
View Configuration Guide →
MiniMax
Ultra-long text large model launched by MiniMax. Supports abab6.5-chat, abab6.5s-chat and other models.
View Configuration Guide →
StepFun
Step series models focused on long context. Supports step-1-8k, step-1-32k and other models.
View Configuration Guide →
SenseNova
SenseNova series models launched by SenseTime. Supports SenseChat-5, SenseChat-Turbo and other models.
View Configuration Guide →
Baichuan AI
Baichuan series models launched by Baichuan Intelligence. Supports Baichuan4, Baichuan3-Turbo, Baichuan3-Turbo-128k and other models.
View Configuration Guide →Local/Private Deployment
Local LLM
Large model service supporting local deployment. Completely private with controllable data security.
View Configuration Guide →
Ollama
Open source tool for running large models locally. Supports DeepSeek-R1, Llama, Qwen and many other models.
View Configuration Guide →
vLLM
High-performance large model inference engine. Supports PagedAttention technology with throughput improvement up to 24x.
View Configuration Guide →
Xorbits Inference
Inference framework supporting multiple models. One-click deployment, supports 100+ open source models.
View Configuration Guide →
Regolo
Enterprise-grade private deployment solution. High availability guarantee with professional technical support.
View Configuration Guide →1. Selection Recommendations
1.1 By Use Case
1.1.1 Technical Interview Scenarios (Algorithms, Programming, Architecture)
Top-Tier Reasoning Capability:
- OpenAI GPT-5: Latest flagship model, strongest overall capability, best technical understanding depth
- Anthropic Claude Sonnet 4.5: World's strongest coding model, excellent reasoning capability
- Google Gemini 2.0: Strong multimodal capability, good latest technology support
- Zhipu GLM-4: One of the strongest technical models in China, excellent Chinese understanding
- DeepSeek-R1: Clear reasoning path, especially suitable for algorithm problems
- SenseNova SenseNova-V6-5-Pro: Latest flagship model, powerful capabilities
High Value Options:
- DeepSeek-V3: High value, technical capability matches flagship models
- Baichuan AI Baichuan4: Accurate technical understanding, fast response
- Tencent Cloud DeepSeek-V3: Good stability, enterprise-grade support
1.1.2 Code Generation and Programming Assistance
Professional Code Models:
- Anthropic Claude Sonnet 4.5: World's strongest coding model, top-tier code understanding and generation capability
- SenseNova Qwen3-Coder: Qwen's latest code model, excellent programming capability
- SenseNova Qwen2-5-Coder: High code generation accuracy, supports multiple programming languages
- OpenAI GPT-5: Latest flagship model, powerful code capability
- MiniMax MiniMax-M2: Optimized for coding tasks and Agent workflows
Fast Response:
- Baichuan AI Baichuan3-Turbo: Fast code generation, high quality
- StepFun step-2-mini: Lightweight model, fast response
1.1.3 Long Document Processing Scenarios
Ultra-Long Context Experts:
- StepFun step-1-256k: Supports 256K tokens ultra-long context, best for ultra-long document processing
- MiniMax abab6.5s-chat: Supports 245K tokens context, excellent long text understanding depth
- SenseNova SenseChat-128K: 128K ultra-long context, suitable for complex document analysis
- Kimi: Ultra-long context understanding, long text processing expert
Long Context Value Options:
- StepFun step-1-32k: 32K context, high value
- Baichuan AI Baichuan3-Turbo-128k: 128K context, moderate cost
- SenseNova SenseChat-32K: 32K context, stable and reliable
- MiniMax abab6.5t-chat: Fast response, suitable for regular long text
1.1.4 Behavioral Interview Scenarios (HR, Communication, Soft Skills)
Excellent Dialogue Understanding:
- OpenAI GPT-4o Mini: Lightweight multimodal model, accurate dialogue understanding
- Kimi: Strong context understanding capability, smooth multi-turn conversations
- DeepSeek-Chat: Natural dialogue, low cost
- SenseNova SenseChat-5: Excellent Chinese dialogue capability, accurate understanding
- MiniMax abab6.5g-chat: General dialogue model, good daily usage experience
Fast Response Options:
- Tencent Hunyuan Hunyuan-Standard: Fast response, high stability
- Baidu Qianfan ERNIE-3.5: Good Chinese understanding, fast
- StepFun step-1-8k: Fast response, high value
- SenseNova SenseChat-Turbo: Fast model, suitable for high-frequency dialogue
1.1.5 Multimodal Scenarios (Image-Text Understanding, Visual Analysis)
Multimodal Capability:
- Google Gemini 2.0: Strongest multimodal capability, supports text, images, video
- Google Gemini 1.5 Pro: Long context multimodal support
- SenseNova SenseNova-V6-5-Omni: Full-modal interaction, strong real-time dialogue capability
- SenseNova SenseChat-Vision: Excellent visual understanding capability, smooth image-text dialogue
1.1.6 Special Scenarios
Cantonese Dialogue:
- SenseNova SenseChat-5-Cantonese: Cantonese dialogue expert, accurate dialect understanding
Role Playing:
- SenseNova SenseChat-Character-Pro: Advanced role-playing capability
- SenseNova SenseChat-Character: Basic role playing
Reasoning Chain:
- SenseNova SenseNova-V6-Reasoner: Reasoning task expert, deep logical analysis
- DeepSeek-R1: Clear reasoning path, visible thinking process
Agent Workflows:
- MiniMax MiniMax-M2: Designed for Agent workflows, excellent coding tasks
1.1.7 Data Security and Privacy Scenarios
Complete Private Deployment:
- Local LLM: Data completely stays on premises, absolute control
- Ollama: Open source local running, supports DeepSeek-R1, Llama, Qwen and many other models
- vLLM: High-performance inference engine, throughput improvement up to 24x
- Xorbits Inference: One-click deployment, supports 100+ open source models
- Regolo: Enterprise-grade private solution, professional technical support
Enterprise-Grade Cloud Services:
- Azure OpenAI: Microsoft cloud platform, strong enterprise compliance
- Amazon Bedrock: AWS platform, multiple model choices
- Alibaba Cloud Bailian: Chinese enterprise platform, Qwen series models
- Tencent Cloud: Enterprise-grade support, high stability
1.2 By Budget
1.2.1 High Budget (Pursuing Ultimate Performance)
International Top-Tier Models:
- OpenAI GPT-5: Latest flagship model, strongest overall capability, suitable for high-value scenarios
- Anthropic Claude Sonnet 4.5: World's strongest coding model
- Anthropic Claude Opus 4.1: Top-tier reasoning capability, deep analysis
- Google Gemini 2.0: Strongest multimodal capability
Chinese Flagship Models:
- Zhipu GLM-4: One of the strongest technical models in China
- SenseNova SenseNova-V6-5-Pro: Latest flagship, powerful capabilities
- MiniMax MiniMax-M2: Top-tier for coding and Agent tasks
1.2.2 Medium Budget (High Value Options)
High Value International Models:
- Anthropic Claude Haiku 4.5: High value, fast and low cost
- OpenAI GPT-5 Mini: Lightweight GPT-5, high value
- Google Gemini 1.5 Pro: Long context, moderate price
High Value Chinese Models:
- DeepSeek-V3: Capability close to top-tier, extremely low price
- Tencent Cloud DeepSeek-V3: Good stability, low cost
- Zhipu GLM-3: Good results, reasonable price
- Kimi: Ultra-long context, high value
- Baichuan AI Baichuan4: Strong technical capability, moderate price
- SenseNova SenseNova-V6-5-Turbo: High-performance fast model
- MiniMax abab6.5s-chat: Ultra-long context, reasonable price
1.2.3 Low Budget (Free Credits/Low Cost)
With Free Credits:
- DeepSeek: Offers free credits, extremely low cost
- SiliconFlow: Focused on AI inference acceleration, multiple open source models
- iFlytek Spark: Offers free trial credits
- Baidu Qianfan: New users have free credits
- Tencent Hunyuan: Offers free trial
- SenseNova: Individual users have free credits after real-name verification
- MiniMax: New user gift package launched in 2025
Fast Lightweight Models:
- OpenAI GPT-4o Mini: Lightweight multimodal model, high value
- StepFun step-1-8k: Fast response, low cost
- StepFun step-2-mini: Latest lightweight, extremely high value
- MiniMax abab6.5t-chat: Fast model, low cost
- SenseNova SenseChat-Turbo: Fast response, high value
- Baichuan AI Baichuan3-Turbo: Fast and stable, friendly price
Completely Free (Local Deployment):
- Ollama: Completely free, supports multiple open source models
- vLLM: Open source inference engine, high performance
- Xorbits Inference: Open source framework, supports 100+ models
1.3 By Region and Network
1.3.1 International Users or Need International Service Access
Prefer International Providers:
- OpenAI: World's strongest AI provider
- Anthropic: Claude series, high security
- Google Gemini: Strong multimodal capability
- Azure OpenAI: Enterprise-grade, global deployment
- Amazon Bedrock: AWS platform, globally available
1.3.2 Chinese Users or Network Restrictions
Chinese Provider Options:
- Zhipu AI: Tsinghua-affiliated, strong technical capability
- DeepSeek: Highest value, strong capability
- Alibaba Cloud Bailian: Enterprise-grade, Qwen series
- Tencent Cloud: Good stability, multiple model choices
- Baidu Qianfan: ERNIE series, excellent Chinese
- iFlytek Spark: Strong speech technology, multiple version choices
- Volcengine: ByteDance, Doubao large model
- Kimi: Moonshot AI, ultra-long context
- Tencent Hunyuan: Tencent self-developed, stable and reliable
Emerging Providers (2025 Recommendations):
- SenseNova: 22 models available, including SenseNova, SenseChat series and third-party models (Qwen, DeepSeek, Kimi)
- Baichuan AI: Baichuan4 strong technical capability, Baichuan3-Turbo-128k good long context support
- MiniMax: Ultra-long text expert, abab6.5s-chat supports 245K tokens
- StepFun: step-1-256k supports 256K tokens ultra-long context, API fully compatible with OpenAI
1.4 Quick Selection Guide
| Scenario | First Choice | Alternatives | Budget Option |
|---|---|---|---|
| Technical Interview | OpenAI GPT-5 | Claude Sonnet 4.5, Zhipu GLM-4 | DeepSeek-V3 |
| Code Generation | Claude Sonnet 4.5 | SenseNova Qwen3-Coder, OpenAI GPT-5 | Baichuan AI Baichuan3-Turbo |
| Long Document Processing | StepFun step-1-256k | MiniMax abab6.5s-chat, SenseNova SenseChat-128K | Kimi |
| Behavioral Interview | GPT-4o Mini | Kimi, SenseNova SenseChat-5 | DeepSeek-Chat |
| Multimodal | Google Gemini 2.0 | SenseNova SenseNova-V6-5-Omni | SenseNova SenseChat-Vision |
| Data Security | Ollama | vLLM, Regolo | Local LLM |
| Ultra-Long Context | StepFun step-1-256k | MiniMax abab6.5s-chat | Baichuan AI Baichuan3-Turbo-128k |
| Fast Response | StepFun step-2-mini | SenseNova SenseChat-Turbo | MiniMax abab6.5t-chat |
| Agent Workflows | MiniMax MiniMax-M2 | Claude Sonnet 4.5, OpenAI GPT-5 | DeepSeek-V3 |
| Reasoning Tasks | DeepSeek-R1 | Claude Opus 4.1, SenseNova SenseNova-V6-Reasoner | Zhipu GLM-4 |