Skip to content
ajan.io

Docs menu

On this page

Recommended Models

Guidelines for selecting models in BYOK workspaces

Which models to add

AI models evolve quickly. We recommend adding at least one flagship model from each major provider. This gives your users access to the best available model for different tasks while keeping your setup manageable.

When a provider releases a new model, add it alongside the existing one rather than replacing it. Users may have agents or workflows that rely on specific models, so removing them without notice can cause disruption.

Select a provider to see the recommended model types and their configuration values.

Model-specific configuration

Use these values when configuring models manually. With a prebuilt ajan config, these are applied automatically.

Add these model types:

  • Latest flagship - Most capable for complex tasks
  • Efficient variant (mini/nano) - Fast, cost-effective for everyday use
  • Reasoning model (o-series) - For analytical and mathematical tasks

GPT-5.5

ModelAPI TypeContext SizeMaxOutput TokensConfiguration
GPT-5.5Responses API200,00016,000Enable Always show reasoning, reasoning: none, verbosity: low

GPT-5.4

ModelAPI TypeContext SizeMaxOutput TokensConfiguration
GPT-5.4Responses API200,00032,000Reasoning: none, verbosity: medium
GPT-5.4 ThinkingResponses API200,00032,000Enable Always show reasoning, reasoning: medium, verbosity: medium
GPT-5.4 MiniResponses API200,00032,000None

GPT-5.2

ModelAPI TypeContext SizeMaxOutput TokensConfiguration
GPT-5.2Responses API200,00032,000Reasoning: none, verbosity: medium
GPT-5.2 ThinkingResponses API200,00032,000Enable Always show reasoning, reasoning: medium, verbosity: medium
GPT-5.2 ProResponses API200,00016,384Enable Always show reasoning, reasoning: medium, verbosity: low

GPT-5.1

ModelAPI TypeContext SizeMaxOutput TokensConfiguration
GPT-5.1Responses API200,00032,000Reasoning: medium, verbosity: medium
GPT-5.1 ThinkingResponses API200,00032,000Enable Always show reasoning, reasoning: medium, verbosity: medium
GPT-5.1 Thinking FastResponses API200,00032,000Enable Always show reasoning, reasoning: none, verbosity: low

GPT-5

ModelAPI TypeContext SizeMaxOutput TokensConfiguration
GPT-5Responses API200,00032,000Enable Always show reasoning, reasoning: minimal, verbosity: low
GPT-5 ThinkingResponses API200,00032,000Enable Always show reasoning
GPT-5 ProResponses API200,00016,384Enable Always show reasoning, reasoning: high, verbosity: low
GPT-5 MiniResponses API200,00032,000Reasoning: minimal, verbosity: low
GPT-5 NanoResponses API200,00032,000Reasoning: minimal, verbosity: low

GPT-4.1

ModelAPI TypeContext SizeMaxOutput TokensConfiguration
GPT-4.1Completion API200,00032,768Good for data analysis
GPT-4.1 MiniCompletion API200,00032,768None
GPT-4.1 NanoCompletion API200,00032,768None

GPT-4o

ModelAPI TypeContext SizeMaxOutput TokensConfiguration
GPT-4oCompletion API128,00016,384None
GPT-4o MiniCompletion API128,00016,384None

GPT oss

ModelAPI TypeContext SizeMaxOutput TokensConfiguration
GPT oss (120b)Completion API128,00016,384Enable Always show reasoning

Reasoning models (o-series)

ModelAPI TypeContext SizeMaxOutput TokensConfiguration
o3Responses API200,00032,000Enable Always show reasoning
o3 Mini highResponses API200,00032,000Enable Always show reasoning, reasoning: high
o3 ProResponses API200,00032,000Enable Always show reasoning
o4 MiniResponses API200,00032,000Enable Always show reasoning

Add these model types:

  • Sonnet - Balanced model for most tasks, excellent at coding and writing
  • Sonnet Reasoning - Enhanced version for complex logical tasks
  • Opus (if available) - Most intelligent for demanding tasks

Claude Opus

ModelAPI TypeContext SizeMaxOutput TokensConfiguration
Claude Opus 4.7 1MCompletion API1,000,00032,000None
Claude Opus 4.7Completion API200,00032,000None
Claude Opus 4.6 1MCompletion API1,000,00032,000None
Claude Opus 4.6Completion API200,00032,000None
Claude Opus 4.6 ReasoningCompletion API200,00032,000Enable Always show reasoning
Claude Opus 4.5Completion API200,00032,000None
Claude Opus 4.5 ReasoningCompletion API200,00032,000Enable Always show reasoning

Claude Sonnet

ModelAPI TypeContext SizeMaxOutput TokensConfiguration
Claude Sonnet 4.6Completion API200,00032,000None
Claude Sonnet 4.6 ReasoningCompletion API200,00032,000Enable Always show reasoning
Claude Sonnet 4.5Completion API200,00032,000None
Claude Sonnet 4.5 ReasoningCompletion API200,00032,000Enable Always show reasoning
Claude Sonnet 4Completion API200,00032,000None
Claude Sonnet 3.7Completion API200,00032,000None
Claude Sonnet 3.7 ReasoningCompletion API200,00032,000Enable Always show reasoning
Claude Sonnet 3.5Completion API200,0008,000None

Claude Haiku

ModelAPI TypeContext SizeMaxOutput TokensConfiguration
Claude Haiku 4.5Completion API200,00032,000None

Add these model types:

  • Gemini Pro - Flagship model with advanced multimodal capabilities
  • Gemini Flash - Fast, efficient for real-time applications

Gemini Pro

ModelAPI TypeContext SizeMaxOutput TokensConfiguration
Gemini 3.1 Pro PreviewCompletion API200,00032,000None
Gemini 2.5 ProCompletion API200,00032,000None
Gemini 2.5 Pro ReasoningCompletion API200,00032,000Enable Always show reasoning

Gemini Flash

ModelAPI TypeContext SizeMaxOutput TokensConfiguration
Gemini 3 Flash PreviewCompletion API200,00032,000None
Gemini 2.5 FlashCompletion API200,00032,000None

Mistral

ModelAPI TypeContext SizeMaxOutput TokensConfiguration
Mistral Large 3Completion API200,00016,000None
Mistral Large 2411Completion API128,00016,000None
Mistral MediumCompletion API128,0008,000None

Meta

ModelAPI TypeContext SizeMaxOutput TokensConfiguration
Llama 4 MaverickCompletion API200,00016,000None
Llama 3.3 70BCompletion API128,00016,000None

DeepSeek

ModelAPI TypeContext SizeMaxOutput TokensConfiguration
DeepSeek r1Completion API128,0008,000Enable Always show reasoning
DeepSeek v3.1Completion API128,00016,000None