Skip to main content

Supported Models

All models available through the Bankr LLM Gateway. Use bankr llm models to fetch the live list.

Discounts may apply

Time-bounded model discounts are applied automatically when you're eligible. Check bankr.bot/llm for active promos and your effective pricing.

Claude (Anthropic)​

ModelContextInputBest For
claude-fable-5.11Mtext, imageMost powerful Claude, long-running autonomous work
claude-fable-51Mtext, imagePrevious generation Fable
claude-opus-51Mtext, imageLatest frontier, advanced reasoning
claude-opus-4.81Mtext, imagePrevious generation Opus
claude-opus-4.71Mtext, imageAdvanced reasoning, analysis
claude-opus-4.61Mtext, imagePrevious generation Opus
claude-opus-4.5200Ktext, imagePrevious generation Opus
claude-sonnet-51Mtext, imageLatest Sonnet, balanced speed and quality
claude-sonnet-4.61Mtext, imagePrevious generation Sonnet
claude-sonnet-4.51Mtext, imagePrevious generation Sonnet
claude-haiku-4.5200Ktext, imageFast, cost-effective
Claude Code and the [1m] suffix

Claude Code exposes the 1M-token tier as a [1m] model suffix (claude-opus-5[1m]). It's optional here — the gateway strips it and enables the 1M beta from the context window above — but if you use it, quote it: zsh reads [1m] as a glob and errors with no matches found. See Claude Code.

Gemini (Google)​

ModelContextInputBest For
gemini-3.8-flash1Mtext, imageLatest Flash — coding and agents
gemini-3.7-flash1Mtext, imagePrevious Flash — coding and agents
gemini-3.6-flash1Mtext, imageFast, cost-effective multimodal
gemini-3.5-flash-lite——Deprecated — use gemini-3.8-flash
gemini-3.5-flash1Mtext, imageFast, cost-effective multimodal
gemini-3.1-pro1Mtext, imageLatest frontier Gemini reasoning
gemini-3.1-flash-lite——Deprecated — use gemini-3.8-flash
gemini-3-pro——Deprecated — use gemini-3.1-pro
gemini-3-flash1Mtext, imageFast, high throughput
gemini-2.5-pro——Deprecated — use gemini-3.1-pro
gemini-2.5-flash——Deprecated — use gemini-3.8-flash

Gemma (Google)​

Open-weight Google models routed through Vertex AI (primary) with OpenRouter fallback.

ModelContextInputBest For
gemma-4-31b-it262Ktext, imageOpen-weight, multimodal
gemma-4-26b-a4b-it262Ktext, imageOpen-weight, lightweight MoE

GPT (OpenAI)​

ModelContextInputBest For
gpt-6-astra1Mtext, imageLatest frontier OpenAI reasoning
gpt-5.6-sol1Mtext, imagePrevious frontier OpenAI flagship
gpt-5.6-terra1Mtext, imageBalanced quality and cost
gpt-5.6-luna1Mtext, imageFast, high-volume work
gpt-5.51Mtext, imagePrevious frontier, native caching
gpt-5.41Mtext, imageFrontier OpenAI reasoning
gpt-5.4-mini400Ktext, imageFast, economical
gpt-5.4-nano400Ktext, imageUltra-fast, lowest cost
gpt-5.2400KtextPrevious generation reasoning
gpt-5.2-codex400KtextCode generation
gpt-5-mini400KtextFast, economical
gpt-5-nano400KtextUltra-fast, lowest cost

Grok (xAI)​

ModelContextInputBest For
grok-4.6500Ktext, imageLatest xAI reasoning
grok-4.5500Ktext, imagePrevious xAI flagship
grok-4.31Mtext, imageValue-tier xAI reasoning
grok-4.202Mtext, imagexAI frontier reasoning
grok-4.1-fast——Deprecated — use grok-4.3

Kimi (Moonshot AI)​

ModelContextInputBest For
kimi-k31Mtext, imageFlagship Moonshot, multimodal
kimi-k2.7-code262KtextMoonshot coding/agentic
kimi-k2.6262KtextMoonshot reasoning
kimi-k2.5262KtextLong-context reasoning

Qwen (Alibaba)​

ModelContextInputBest For
qwen3.8-max1Mtext, imageFlagship Qwen, multimodal
qwen3.8-flash1Mtext, imageLatest Qwen flash, multimodal
qwen3.7-max1MtextPrevious flagship, reasoning
qwen3.7-plus1MtextAgentic reasoning, lower cost
qwen3.7-flash1Mtext, imagePrevious flash, vision
qwen3.6-flash1MtextFast, economical
qwen3-coder262KtextCode generation, debugging
qwen3.5-plus1MtextLong-context reasoning
qwen3.5-flash1MtextFast, economical

DeepSeek​

ModelContextInputBest For
deepseek-v4.1-flash1Mtext, imageMid-tier Flash — agents, coding, vision
deepseek-v4-pro-08131MtextFrontier, high-capacity reasoning
deepseek-v4-pro1MtextPrevious V4 Pro build (0423)
deepseek-v4-flash1MtextBudget Flash — fast, cost-effective
deepseek-v3.2163KtextPrevious generation reasoning

GLM (Z.ai)​

ModelContextInputBest For
glm-5.3-flashx1Mtext, imageHigh-speed Flash — ~200 tok/s
glm-5.3-flash1Mtext, imageEfficient coding and agents
glm-5.31MtextLatest Z.ai flagship, 1M context
glm-5.21MtextPrevious Z.ai flagship
glm-5.1202KtextZ.ai reasoning
glm-5202KtextZ.ai reasoning
glm-5-turbo202KtextFast Z.ai

MiniMax​

ModelContextInputBest For
minimax-m3524Ktext, imageFlagship MSA architecture, multimodal
minimax-m2.7204KtextPrevious generation reasoning
minimax-m2.7-highspeed204KtextHigh-throughput MiniMax
minimax-m2.5204KtextLong-context reasoning