Skip to main content

​
Overview

SaveGate provides access to 50+ state-of-the-art AI models from leading providers. All models are accessible through a unified API using the same authentication and request format.

​
OpenAI Models

​
GPT-4 Family

GPT-4

Model ID: gpt-4
  • Most capable GPT-4 model
  • Best for complex reasoning
  • Context: 8K tokens

GPT-4 Turbo

Model ID: gpt-4-turbo, gpt-4-turbo-preview
  • Faster and cheaper than GPT-4
  • Context: 128K tokens
  • JSON mode support

GPT-4o

Model ID: gpt-4o, gpt-4o-mini
  • Multimodal (text + vision)
  • Fast and efficient
  • Context: 128K tokens

O1 Series

Model IDs: o1-preview, o1-mini
  • Advanced reasoning
  • Best for complex problems
  • Extended thinking time

​
GPT-3.5 Family

Model IDContextBest For
gpt-3.5-turbo16KGeneral tasks, cost-effective
gpt-3.5-turbo-16k16KLonger conversations

​
Anthropic Models

​
Claude 3 Family

Claude 3.5 Sonnet

Model ID: claude-3-5-sonnet-20241022
  • Latest and most capable
  • Best balance of speed/capability
  • Context: 200K tokens
  • Excellent for coding

Claude 3 Opus

Model ID: claude-3-opus-20240229
  • Most capable Claude model
  • Best for complex tasks
  • Context: 200K tokens

Claude 3 Sonnet

Model ID: claude-3-sonnet-20240229
  • Balanced performance
  • Good for most tasks
  • Context: 200K tokens

Claude 3 Haiku

Model ID: claude-3-haiku-20240307
  • Fastest Claude model
  • Most cost-effective
  • Context: 200K tokens

​
Google Models

​
Gemini Family

Gemini 1.5 Pro

Model ID: gemini-1.5-pro
  • Multimodal capabilities
  • Context: 1M tokens
  • Best for long context

Gemini 1.5 Flash

Model ID: gemini-1.5-flash
  • Fast and efficient
  • Context: 1M tokens
  • Cost-effective

Gemini Pro

Model ID: gemini-pro
  • General purpose
  • Good performance
  • Reliable choice

Gemini Pro Vision

Model ID: gemini-pro-vision
  • Multimodal (text + images)
  • Vision understanding
  • Creative tasks

​
Meta Models

​
Llama 3 Family

Model IDSizeContextBest For
meta-llama/llama-3-70b-instruct70B8KGeneral tasks, reasoning
meta-llama/llama-3-8b-instruct8B8KFast, cost-effective
meta-llama/llama-2-70b-chat70B4KConversations
meta-llama/llama-2-13b-chat13B4KLightweight chat

​
Mistral Models

Mistral Large

Model ID: mistral-large-latest
  • Most capable Mistral model
  • Multilingual support
  • Function calling

Mistral Medium

Model ID: mistral-medium-latest
  • Balanced performance
  • Good for most tasks
  • Cost-effective

Mistral Small

Model ID: mistral-small-latest
  • Fast and lightweight
  • Simple tasks
  • Most affordable

Mixtral 8x7B

Model ID: mixtral-8x7b-instruct
  • Mixture of Experts
  • Strong performance
  • Open weights

​
Model Selection Guide

​
By Use Case

Best Models:
  1. claude-3-5-sonnet-20241022 - Best overall
  2. gpt-4-turbo - Strong alternative
  3. claude-3-opus-20240229 - Complex problems
Why: Advanced reasoning, code understanding, and generation

​
By Budget

  • o1-preview - Advanced reasoning
  • claude-3-opus-20240229 - Most capable Claude
  • gpt-4 - Original GPT-4
Use when: Quality is critical, complex tasks
  • claude-3-5-sonnet-20241022 - Best overall value
  • gpt-4-turbo - Fast and capable
  • gemini-1.5-pro - Long context
Use when: Need quality at reasonable cost
  • gpt-3.5-turbo - Fast and cheap
  • claude-3-haiku-20240307 - Fastest Claude
  • gemini-1.5-flash - Fast Gemini
  • mistral-small-latest - Affordable
Use when: Simple tasks, high volume

​
Model Capabilities

​
Function Calling

Models that support function/tool calling:
  • ✅ All GPT-4 models
  • ✅ GPT-3.5 Turbo
  • ✅ Claude 3 family
  • ✅ Mistral Large
  • ✅ Gemini Pro

​
Vision (Multimodal)

Models that support image understanding:
  • ✅ GPT-4o, GPT-4o Mini
  • ✅ Claude 3 family (all models)
  • ✅ Gemini Pro Vision
  • ✅ Gemini 1.5 Pro

​
JSON Mode

Models with guaranteed JSON output:
  • ✅ GPT-4 Turbo
  • ✅ GPT-3.5 Turbo
  • ✅ Gemini 1.5 Pro

​
Using Models

Simply specify the model ID in your API request: 
response = client.chat.completions.create(
    model="claude-3-5-sonnet-20241022",  # Specify model here
    messages=[{"role": "user", "content": "Hello!"}]
)

​
Model Updates

SaveGate automatically updates model versions to the latest stable releases:
  • gpt-4 → Latest GPT-4 version
  • claude-3-5-sonnet-latest → Latest Claude 3.5 Sonnet
  • gemini-pro → Latest Gemini Pro
For version pinning, use specific model IDs with dates (e.g., claude-3-5-sonnet-20241022).
Check the pricing page for detailed cost information for each model.