Max Mode

Override the default agent LLM with a frontier, flagship or balanced model from the gateway.

By default, the Bankr agent uses Gemini 3.8 Flash — a fast, cost-effective model. Max Mode lets you swap it for a more capable model from the LLM Gateway (Claude Opus 5, GPT-5.6, Gemini 3.1 Pro, etc.) so the agent uses a stronger model for complex tasks. Only the frontier, flagship and balanced lines are offered; light models such as Claude Haiku or GPT Mini are available through the gateway API but not as a Max Mode choice.

Usage is billed per-token from your LLM credit balance.

Credits Required

Max Mode requires LLM credits. Top up at bankr.bot/llm?tab=credits before enabling, or your messages will fail.

How It Works

1. You enable Max Mode and pick a model
2. Every agent message uses that model instead of the default
3. Token usage (input + output) is tracked and billed at the model's rate
4. A usage badge appears below each response showing the model, token count, and cost

Your Max Mode setting is saved to your wallet and applies across all platforms — web terminal, Farcaster, Twitter/X, Telegram, and automations.

Using Max Mode in the Web Terminal

1. Go to bankr.bot and open the chat
2. Click the Max button above the input box to enable it (turns purple when active)
3. Click the model name next to it to open the model picker
4. Search or filter by provider, then select a model
5. Start chatting — the agent will use your chosen model

To disable, click the Max button again. Your model selection is remembered for next time.

Info Icon

Hover the ⓘ icon next to the Max button for a quick explainer.

Using Max Mode in the CLI

Pass the --model (or -m) flag to any prompt:

bankr "analyze my portfolio" --model claude-opus-5

bankr "what are the top memecoins today?" -m gemini-3.1-pro

You can combine it with thread continuation:

bankr "tell me more" --continue --model claude-sonnet-4.6

Available CLI Models

bankr llm models    # List all available models with pricing

A selection of popular --model values (run bankr llm models for the full, current list):

Model	Provider	Input/M	Output/M
claude-fable-5.1	Anthropic	$10.00	$50.00
claude-fable-5	Anthropic	$10.00	$50.00
claude-opus-5	Anthropic	$5.00	$25.00
claude-opus-4.8	Anthropic	$5.00	$25.00
claude-sonnet-5	Anthropic	$2.00	$10.00
claude-sonnet-4.6	Anthropic	$3.00	$15.00
gpt-6-astra	OpenAI	$10.00	$50.00
gpt-5.6-sol	OpenAI	$4.00	$20.00
gemini-3.1-pro	Google	$2.00	$12.00
grok-4.6	xAI	$2.00	$6.00
grok-4.5	xAI	$2.00	$6.00
kimi-k3	Moonshot	$3.00	$15.00
qwen3.8-max	Alibaba	$2.00	$6.00

The web terminal, X, Telegram, automations, the API and the CLI all offer the same ranks. The picker and the CLI list the two newest releases of every frontier, flagship and balanced line; the API also accepts an older release of those lines while the gateway still serves it.

Pricing

Each model has its own per-million-token rate for input and output. You can see pricing in:

• The model picker modal in the web terminal (shown next to each model)
• The Models tab at bankr.bot/llm
• The CLI via bankr llm models

Cost is calculated as:

cost = (input_tokens / 1,000,000) × input_rate
     + (output_tokens / 1,000,000) × output_rate

After each Max Mode response, a usage badge shows the exact token count and cost.

Managing Credits

bankr llm credits                  # Check your balance
bankr llm credits add 25           # Add $25 (USDC)
bankr llm credits auto --enable    # Enable auto top-up

Or manage credits in the browser at bankr.bot/llm?tab=credits, where you can also send credits to another Bankr user — or just ask the agent: "send $20 of my LLM credits to @username".

See LLM Gateway Overview for full credit management docs, including sending credits.

FAQ

Do I need Bankr Club to use Max Mode? No. Max Mode is an alternative to Bankr Club — either one unlocks the agent. You can use both together.

Does Max Mode work with automations? Yes. Your Max Mode setting applies globally, including scheduled automations and social platform interactions. An automation that skips three runs in a row for lack of credits is paused, and you're notified once — top up and resume it from the terminal.

What happens if I run out of credits? It depends when. If you're out of credits when you send a message, the agent replies asking you to top up rather than answering on the default model — the one exception is a message you send on X, where it quietly falls back to the standard model instead of posting a public top-up reply (automations running on X take the pause path above). If credits run out partway through a turn, the agent stops there, tells you what it spent, and saves its progress in the thread so you can top up and ask it to continue. Enable auto top-up to avoid the interruption.

What happens if I hit my daily spend budget? Different problem, different fix — topping up won't help. The budget is an optional cap on how much the whole account can spend on LLMs in any rolling 24-hour window, covering Max Mode runs as well as gateway requests from your API keys. When a run hits it, the agent stops where it is, saves its progress in the thread, and tells you the budget is what stopped it. Capacity comes back gradually as earlier usage ages past 24 hours old — there's no midnight reset — or you can raise the cap in the Settings tab at bankr.bot/llm, where the card also shows what the window has spent so far.

Can I change models mid-conversation? Yes. Each message uses whatever model is active when you send it. You can switch models between messages freely.