← Back to Home

Supported Models

Juggernaut supports multiple LLM providers. Models are referenced using the format: `[provider]-[model_name]` (e.g., openai-gpt-4, google-gemini-2.5-pro).

Model Naming Convention

Critical: Always use the full qualified name format provider-model. Even if model names appear duplicated across providers, this prefix ensures correct routing.

Examples:

  • Correct: openai-gpt-4, anthropic-claude-3-5-sonnet, google-gemini-2.5-pro
  • Incorrect: gpt-4, claude-sonnet, gemini-pro

OpenAI

Provider Prefix: openai-

| Model ID | Input Price | Output Price | Description | Modalities | Web Search | |----------|-------------|--------------|-------------|------------|------------| | openai-gpt-5.2-pro | $21.00 | $168.00 | Most capable GPT-5 reasoning model | text → text, json | ✅ | | openai-gpt-5.2 | $1.75 | $14.00 | Standard GPT-5 model | text → text, json | ✅ | | openai-gpt-5-nano | $0.05 | $0.40 | Fast, cost-effective GPT-5 | text → text | ✅ | | openai-gpt-5-mini | $0.25 | $2.00 | Balanced GPT-5 model | text → text, json | ❌ | | openai-gpt-5 | $0.15 | $0.60 | Base GPT-5 model | text → text, json | ✅ | | openai-gpt-4o-2024-08-06 | $2.50 | $10.00 | GPT-4o latest | text, json → text, json | ✅ | | openai-gpt-4o-mini-2024-07-18 | $0.15 | $0.60 | Lightweight GPT-4o | text, json → text, json | ✅ | | openai-gpt-4.1-2025-04-14 | $2.00 | $8.00 | GPT-4.1 standard | text, json → text, json | ✅ | | openai-gpt-4.1-mini-2025-04-14 | $0.40 | $1.60 | Mini variant | text, json → text, json | ✅ | | openai-gpt-4.1-nano-2025-04-14 | $0.10 | $0.40 | Nano variant | text, json → text, json | ✅ | | openai-gpt-4.5-preview-2025-02-27 | $75.00 | $150.00 | Experimental preview | text, json → text, json | ✅ | | openai-gpt-image-1 | $0.02 | $0.02 | Image generation | text, json, image → image | ❌ | | openai-dall-e-3 | $0.02 | $0.02 | DALL-E image synthesis | text, json, image → image | ❌ | | openai-tts-1 | $15.00 | $15.00 | Text-to-speech standard | text, json → speech | ❌ | | openai-tts-1-hd | $30.00 | $30.00 | High-definition TTS | text, json → speech | ❌ | | openai-gpt-4o-mini-tts | $0.60 | $12.00 | GPT-4o powered TTS | text, json → speech | ❌ |

Prices per 1M tokens (or per 1000 for image/TTS models)

Anthropic

Provider Prefix: anthropic-

| Model ID | Input Price | Output Price | Description | Modalities | Web Search | |----------|-------------|--------------|-------------|------------|------------| | anthropic-claude-opus-4-20250514 | $30.00 | $75.00 | Highest capability Claude | text, json → text, json | ✅ | | anthropic-claude-opus-4-1-20250805 | $30.00 | $75.00 | Opus 4.1 variant | text, json → text, json | ✅ | | anthropic-claude-sonnet-4-5-20250929 | $3.00 | $15.00 | Advanced reasoning | text, json → text, json | ✅ | | anthropic-claude-sonnet-4-20250514 | $3.00 | $15.00 | Sonnet standard | text, json → text, json | ✅ | | anthropic-claude-3-7-sonnet | $3.00 | $15.00 | Claude 3.7 | text, json → text, json | ✅ | | anthropic-claude-3-5-sonnet | $3.00 | $15.00 | Claude 3.5 | text, json → text, json | ✅ | | anthropic-claude-3-5-haiku | $0.80 | $4.00 | Fast haiku model | text, json → text, json | ✅ | | anthropic-claude-3-0-opus | $15.00 | $75.00 | Legacy opus | text, json → text, json | ✅ | | anthropic-claude-3-0-haiku | $0.25 | $1.25 | Legacy haiku | text, json → text, json | ✅ |

Google (Gemini)

Provider Prefix: google-

| Model ID | Input Price | Output Price | Description | Modalities | Web Search | |----------|-------------|--------------|-------------|------------|------------| | google-gemini-3-pro-preview | $2.00 | $12.00 | Gemini 3 Pro | text → text, json | ✅ | | google-gemini-3-pro-image-preview | $2.00 | $0.134 | Gemini 3 with image output | text → image | ✅ | | google-gemini-3-flash-preview | $0.50 | $3.00 | Fast Gemini 3 | text → text | ✅ | | google-gemini-2-5-pro | $1.25 | $10.00 | Gemini 2.5 Pro | text, json → text, json | ✅ | | google-gemini-2-5-flash | $0.30 | $2.50 | Fast Gemini 2.5 | text, json → text, json | ✅ | | google-gemini-2-5-flash-image | $0.00 | $0.00 | Image-capable flash | text → text | ❌ | | google-gemini-2-5-pro-tts | $0.15 | $0.60 | Text-to-speech | text → speech | ❌ | | google-gemini-2-5-flash-tts | $15.00 | $15.00 | Flash TTS | text, json → speech | ❌ | | google-gemini-2-0 | $0.10 | $0.40 | Gemini 2.0 | text, json → text, json | ✅ |

Groq

Provider Prefix: groq-

High-performance inference for open-source models.

| Model ID | Input Price | Output Price | Description | Modalities | Web Search | |----------|-------------|--------------|-------------|------------|------------| | groq-llama-4-maverick-17b | $0.20 | $0.60 | Llama 4 Maverick | text, json → text, json | ❌ | | groq-llama-4-scout-17b | $0.11 | $0.34 | Llama 4 Scout | text, json → text, json | ❌ | | groq-llama-3-3-70b-versatile | $0.59 | $0.79 | Llama 3.3 70B | text, json → text, json | ❌ | | groq-llama-3-1-8b-instant | $0.05 | $0.08 | Fast 8B model | text, json → text, json | ❌ | | groq-llama-3-70b | $0.59 | $0.79 | Llama 3 70B | text, json → text, json | ❌ | | groq-llama-3-8b | $0.05 | $0.08 | Llama 3 8B | text, json → text, json | ❌ | | groq-llama-guard-3-8b | $0.20 | $0.20 | Safety filter model | text, json → text, json | ❌ | | groq-llama-guard-4-12b | $0.20 | $0.20 | Llama Guard 4 | text, json → text, json | ❌ | | groq-mixtral-8x7b | $0.70 | $0.70 | Mixtral 8x7B MoE | text, json → text, json | ❌ | | groq-mixtral-8x22b | $2.00 | $6.00 | Mixtral 8x22B | text, json → text, json | ❌ | | groq-mistral-large-3 | $2.00 | $6.00 | Mistral Large | text, json → text, json | ✅ | | groq-mistral-medium-3 | $0.40 | $2.00 | Mistral Medium | text, json → text, json | ✅ | | groq-mistral-small-3-1 | $0.10 | $0.30 | Mistral Small | text, json → text, json | ✅ | | groq-mistral-saba | $0.20 | $0.60 | Mistral Saba | text, json → text, json | ✅ | | groq-mistral-saba-24b | $0.79 | $0.79 | 24B variant | text, json → text, json | ❌ | | groq-ministral-3b-24-10 | $0.04 | $0.04 | 3B parameter | text, json → text, json | ✅ | | groq-ministral-8b-24-10 | $0.10 | $0.10 | 8B parameter | text, json → text, json | ✅ | | groq-mistral-nemo | $0.15 | $0.15 | Nemo variant | text, json → text, json | ✅ | | groq-pixtral-12b | $0.15 | $0.15 | Vision model | text, json → text, json | ✅ | | groq-pixtral-large-3 | $2.00 | $6.00 | Large vision | text, json → text, json | ✅ | | groq-gemma-2-9b | $0.20 | $0.20 | Google Gemma | text, json → text, json | ❌ | | groq-qwen-qwq-32b | $0.29 | $0.39 | Qwen reasoning | text, json → text, json | ❌ | | groq-deepseek-r1-distill-llama-70b | $0.75 | $0.99 | DeepSeek distilled | text, json → text, json | ❌ |

Mistral (Native)

Provider Prefix: mistral-

Direct Mistral AI API access.

| Model ID | Input Price | Output Price | Description | Modalities | Web Search | |----------|-------------|--------------|-------------|------------|------------| | mistral-mistral-large-3 | $2.00 | $6.00 | Flagship model | text, json → text, json | ✅ | | mistral-mistral-medium-3 | $0.40 | $2.00 | Balanced | text, json → text, json | ✅ | | mistral-mistral-small-3-1 | $0.10 | $0.30 | Fast inference | text, json → text, json | ✅ | | mistral-mistral-saba | $0.20 | $0.60 | Specialized | text, json → text, json | ✅ | | mistral-ministral-3b-24-10 | $0.04 | $0.04 | Edge deployment | text, json → text, json | ✅ | | mistral-ministral-8b-24-10 | $0.10 | $0.10 | Edge 8B | text, json → text, json | ✅ | | mistral-pixtral-large-3 | $2.00 | $6.00 | Vision + language | text, json → text, json | ✅ | | mistral-pixtral-12b | $0.15 | $0.15 | Small vision | text, json → text, json | ✅ |

Cohere

Provider Prefix: cohere-

Enterprise language models.

| Model ID | Input Price | Output Price | Description | Modalities | Web Search | |----------|-------------|--------------|-------------|------------|------------| | cohere-command-a-03-2025 | $2.50 | $10.00 | Command A | text, json → text, json | ✅ | | cohere-command-r-plus-08-2024 | $2.50 | $10.00 | Command R+ | text, json → text, json | ✅ | | cohere-command-r-08-2024 | $0.15 | $0.60 | Command R | text, json → text, json | ✅ | | cohere-command-r7b-12-2024 | $0.04 | $0.15 | Lightweight 7B | text, json → text, json | ✅ |

DeepSeek

Provider Prefix: deepseek-

Chinese-developed reasoning models.

| Model ID | Input Price | Output Price | Description | Modalities | Web Search | |----------|-------------|--------------|-------------|------------|------------| | deepseek-deepseek-reasoner | $0.28 | $0.42 | Reasoning specialist | text, json → text, json | ✅ | | deepseek-deepseek-chat | $0.28 | $0.42 | General chat | text, json → text, json | ✅ |

Other Providers

| Provider | Model ID | Input | Output | Description | Web Search | |----------|----------|-------|--------|-------------|------------| | Xiaomi | xiaomi-mimo-v2-flash | $0.10 | $0.30 | Xiaomi AI model | ✅ | | ThirdWeb | thirdweb-nebula-t1 | $0.20 | $0.60 | Web3/Blockchain focused | ❌ |

Model Selection Guide

By Use Case

| Task | Recommended Models | Reason | |------|-------------------|--------| | Complex Analysis | anthropic-claude-opus-4, openai-gpt-5.2-pro | Superior reasoning, large context windows | | Fast Iteration | groq-llama-3-1-8b-instant, openai-gpt-5-nano | Low latency, cost-effective | | JSON Generation | openai-gpt-4o, anthropic-claude-3-5-sonnet | Reliable structured output | | Image Generation | openai-gpt-image-1, openai-dall-e-3 | High visual fidelity | | Text-to-Speech | openai-tts-1-hd, google-gemini-2-5-pro-tts | Natural speech synthesis | | Vision Analysis | groq-pixtral-large-3, openai-gpt-4o | Multimodal understanding | | Budget Conscious | deepseek-deepseek-chat, groq-llama-3-8b | High capability, low cost |

Cost Optimization Strategies

  1. Model Tiering: Use expensive models only for complex reasoning steps; use Groq/Llama for bulk processing
  2. Iterative Efficiency: For large arrays, consider batching items rather than individual API calls per element
  3. Cached Context: Re-use analysis outputs across multiple downstream steps rather than re-querying
  4. Fallback Chains: Define defaultModel as mid-tier, override to premium only for specific high-value steps

Important: Always verify modalities compatibility. A model with text → text only cannot handle JSON output schemas without explicit prompting.