AI models we support
Every model available on Kanbu.ai, grouped by how we recommend using it — with price and speed where we have data. Your organization's actual pricing is shown in the app.
Recommended
Best balance of price, performance and latency — our default pick for most support and shopping assistants.
Gemini 3.8 Flash
Our recommendationProvider: GoogleA versatile assistant for helpful, accurate customer conversations.
Connects information clearly, understands detailed product requirements and handles follow-up questions naturally. A great everyday choice for support and shopping advice.
$0.16Average per answer7.0 s · Typical complete answerTypical range: $0.0296 – $0.36GPT-5.6 Luna
Provider: OpenAISmart, economical support with clear answers.
Combines approachable explanations with practical reasoning about prices and product choices. Well suited to everyday support; clear instructions help keep more complex conversations focused.
$0.0476Average per answer6.0 s · Typical complete answerTypical range: $0.0077 – $0.14Gemini 3.5 Flash-Lite
Provider: GoogleQuick, concise answers for busy customer support.
Keeps conversations moving with short response times and direct answers. Ideal for FAQs and routine requests; for intricate calculations or many simultaneous requirements, consider Gemini 3.8 Flash.
$0.0769Average per answer3.7 s · Typical complete answerTypical range: $0.0236 – $0.23Hunyuan HY3
Provider: TencentThoughtful answers at an economical price.
Good at following detailed requirements and explaining product choices without a premium price. A useful choice when answer quality matters more than immediate response speed.
$0.0434Average per answer17.0 s · Typical complete answerTypical range: $0.0125 – $0.0935
Advanced
More depth for advanced, higher-stakes conversations, at a higher cost per answer.
GPT-5.6 Terra
Provider: OpenAIMore depth for demanding conversations.
A premium option for nuanced instructions and detailed explanations. Useful when you want more thorough assistance and can allow a higher cost per answer.
$0.45Average per answer5.0 s · Typical complete answerTypical range: $0.15 – $1.25Claude Sonnet 5
Provider: AnthropicClear, detailed communication for complex requests.
A premium alternative for assistants with extensive instructions and a need for thoughtful explanations. Best suited to conversations where detail is worth the additional cost.
$0.66Average per answer13.7 s · Typical complete answerTypical range: $0.24 – $2.09Grok 4.6
Provider: xAIPremium assistance for challenging questions.
Handles demanding questions with detailed, direct answers. A strong premium option when depth matters; longer searches and complex requests can cost considerably more.
$0.66Average per answer28.6 s · Typical complete answerTypical range: $0.23 – $1.77Muse Spark 1.3
Provider: MetaArticulate support with a natural conversational style.
Offers polished explanations and follows the direction of a conversation well. Suited to more involved customer interactions where a little extra time and cost are acceptable.
$0.36Average per answer17.3 s · Typical complete answerTypical range: $0.0915 – $0.91MiniMax M3
Provider: MiniMaxAn affordable alternative for everyday assistance.
Provides practical answers at a modest cost. Best for clearly defined support tasks; unusually complex or sensitive requests may need a more capable alternative.
$0.0594Average per answer8.0 s · Typical complete answerTypical range: $0.0151 – $0.13
Ultra
The most capable, high-end models available. Usually overkill for a regular support assistant — reach for these only for very difficult workflows, advanced logic and tools, or complex internal use cases.
GPT-5.6 Sol
Provider: OpenAIOpenAI's most capable model, for the hardest reasoning and tool-use tasks.
Handles multi-step reasoning, complex tool orchestration and nuanced logic that other models struggle with. Considerable overkill for everyday support — reserve it for demanding internal workflows or advanced automation.
~$0.54Estimated — not directly measured for this modelGemini 3.1 Pro
Provider: GoogleGoogle's advanced reasoning model for complex, multi-step tasks.
Strong at synthesizing large amounts of context and following intricate instructions. More capability than most support chatbots need — a good fit for sophisticated internal tools or elaborate product logic.
~$0.28Estimated — not directly measured for this modelClaude Opus 5
Provider: AnthropicAnthropic's most capable model, for the most demanding reasoning and agentic tasks.
Best suited to complex multi-step workflows, advanced tool use and nuanced judgment calls. Significant overkill for routine support conversations, but a strong fit for sophisticated internal use cases.
~$0.68Estimated — not directly measured for this model
Experimental
DeepSeek V4.1 Flash
Provider: DeepSeekDetailed reasoning with an economical token price.
A promising choice for thoughtful, detailed answers when speed is less important. Complex searches can take substantially longer or time out, so choose another model for time-sensitive support.
$0.13Average per answer31.1 s · Typical complete answerTypical range: $0.0232 – $0.54StepFun 3.7 Flash
Provider: StepFunAn economical model for exploring new assistant workflows.
Useful for experimenting with everyday conversational tasks. More demanding instructions and tool-heavy requests can need closer supervision; use a recommended model for critical customer workflows.
$0.0554Average per answer10.9 s · Typical complete answerTypical range: $0.0196 – $0.16
Other options
Older and less common models, kept available for existing assistants and specific workflows.
Missing a model? Write to us — we can add it.
gpt-5.5
Provider: OpenAILegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.69Estimated — not directly measured for this modelgpt-5.4
Provider: OpenAILegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.35Estimated — not directly measured for this modelGPT-5.4 Mini
Provider: OpenAILegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.11Estimated — not directly measured for this modelGPT-5.4 Nano
Provider: OpenAILegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.0306Estimated — not directly measured for this modelgpt-4.1
Provider: OpenAILegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.27Estimated — not directly measured for this modelgpt-4.1-mini
Provider: OpenAILegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.0558Estimated — not directly measured for this modelGPT-4.1 Nano
Provider: OpenAILegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.0162Estimated — not directly measured for this modelgpt-5
Provider: OpenAILegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.18Estimated — not directly measured for this modelgpt-5-mini
Provider: OpenAILegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.0386Estimated — not directly measured for this modelgpt-5-nano
Provider: OpenAILegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.0101Estimated — not directly measured for this modelo3-mini-2025-01-31
Provider: OpenAILegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.15Estimated — not directly measured for this modelx-ai/grok-4.5
Provider: xAILegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.26Estimated — not directly measured for this modelx-ai/grok-4-fast
Provider: xAILegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.0286Estimated — not directly measured for this modelx-ai/grok-4.1-fast
Provider: xAILegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.0286Estimated — not directly measured for this modelx-ai/grok-4
Provider: xAILegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.41Estimated — not directly measured for this modelx-ai/grok-4.20
Provider: xAILegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.16Estimated — not directly measured for this modelx-ai/grok-4.3
Provider: xAILegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.16Estimated — not directly measured for this modelgoogle/gemini-2.0-flash-001
Provider: GoogleLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.0228Estimated — not directly measured for this modelgoogle/gemini-2.0-flash-lite-001
Provider: GoogleLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.0129Estimated — not directly measured for this modelgoogle/gemini-2.5-flash
Provider: GoogleLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.046Estimated — not directly measured for this modelGemini 2.5 Flash-Lite
Provider: GoogleLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.0162Estimated — not directly measured for this modelgoogle/gemini-2.5-pro
Provider: GoogleLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.18Estimated — not directly measured for this modelgoogle/gemini-3-flash-preview
Provider: GoogleLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.0716Estimated — not directly measured for this modelGemini 3.1 Flash-Lite
Provider: GoogleLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.0373Estimated — not directly measured for this modelgoogle/gemini-3.1-flash-lite-preview
Provider: GoogleLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.0373Estimated — not directly measured for this modelgoogle/gemini-3.7-flash
Provider: GoogleLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
$0.16Average per answer8.8 s · Typical complete answerTypical range: $0.0567 – $0.38google/gemini-3.6-flash
Provider: GoogleLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.20Estimated — not directly measured for this modelgoogle/gemini-3.5-flash
Provider: GoogleLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.21Estimated — not directly measured for this modelclaude-3-5-sonnet-latest
Provider: AnthropicLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.41Estimated — not directly measured for this modelanthropic/claude-sonnet-4.6
Provider: AnthropicLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.41Estimated — not directly measured for this modelanthropic/claude-opus-4.7
Provider: AnthropicLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.68Estimated — not directly measured for this modelanthropic/claude-opus-4.8
Provider: AnthropicLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.68Estimated — not directly measured for this modelanthropic/claude-opus-4.8-fast
Provider: AnthropicLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$1.35Estimated — not directly measured for this modelanthropic/claude-haiku-4.5
Provider: AnthropicLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
$0.17Average per answer7.6 s · Typical complete answerTypical range: $0.0839 – $0.20z-ai/glm-5.2
Provider: Z.aiLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.18Estimated — not directly measured for this modeldeepseek/deepseek-v4-pro
Provider: DeepSeekLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.22Estimated — not directly measured for this modeldeepseek/deepseek-v4-flash
Provider: DeepSeekLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
$0.0561Average per answer22.9 s · Typical complete answerTypical range: $0.0138 – $0.12xiaomi/mimo-v2.5
Provider: XiaomiLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.0165Estimated — not directly measured for this modelxiaomi/mimo-v2.5-pro
Provider: XiaomiLegacy model retained for existing configurations.
Available for existing assistants and familiar workflows.
~$0.0582Estimated — not directly measured for this model
Approximate usage prices include our standard service margin and conversion into the selected currency. Your organization’s prices are shown in the app. Costs and response times vary with context, tools and answer length. Estimated prices for less common models are derived from typical usage patterns and the model's own published rate, not measured directly.