Industry-leading models.
All in one place.
Choose models across Anthropic, OpenAI, Kimi, DeepSeek, and more — each optimized for different tasks, speeds, and budgets.
AnthropicClaude Fable 5
Anthropic's newest creative model. Excels at long-form narrative, storytelling, and nuanced writing with a distinct personality.
AnthropicClaude Sonnet 5
Next-gen Sonnet with improved reasoning and faster inference. A strong default for most automation workflows.
AnthropicClaude Opus 4.8
Latest Opus generation with enhanced reasoning and better instruction following. Well-suited for multi-step agent tasks and code generation.
AnthropicClaude Opus 4.8 Fast
The fastest Opus variant available. Delivers Opus 4.8 intelligence at significantly lower latency for time-sensitive workflows.
AnthropicClaude Opus 4.7
Iterative improvement on Opus 4.6 with stronger tool use and more reliable structured output. A workhorse for complex agent workflows.
AnthropicClaude Opus 4.6
Anthropic's flagship reasoning model. Handles complex analysis, multi-step planning, and deep research tasks with exceptional accuracy.
AnthropicClaude Sonnet 4.6
Balanced performance and cost. Sonnet handles everyday agent tasks like email drafting, data extraction, and summarization efficiently.
AnthropicClaude Opus 4.6 Fast
Optimized for speed without sacrificing Opus-level quality. Ideal for real-time automation where latency matters.
DeepSeekDeepSeek V4 Pro
DeepSeek's premium model for complex reasoning and code generation. Competes with frontier models at a lower price point.
DeepSeekDeepSeek V4 Flash
Ultra-fast inference at the lowest cost. Designed for high-volume, low-latency tasks like classification and simple extraction.
DeepSeekDeepSeek V3.2
DeepSeek's efficient general-purpose model. Strong performance at a fraction of the cost of Western frontier models.
GoogleGemini 3.1 Pro Preview
Google's top-tier reasoning model. Excels at multi-step analysis, code generation, and tasks requiring large context windows.
GoogleGemini 3 Flash Preview
Google's fast multimodal model. Handles text, images, and long context windows with low latency and aggressive pricing.
GoogleGemma 4 26B A4B
Open-weight model from Google. Solid performance for lighter agent tasks at a very accessible price point.
GoogleGemma 4 31B
Slightly larger open-weight variant with improved instruction following. Good fit for structured extraction and summarization.
MiniMaxMiniMax M3
MiniMax's flagship model with strong multilingual support and creative generation. A competitive option for content-heavy workflows.
Moonshot AIKimi K3
Moonshot's newest generation model. Strong reasoning, tool use, and long-context handling for demanding agent tasks.
Moonshot AIKimi K2.6
Updated Kimi model with improved accuracy and faster inference. Handles multi-step workflows with reliable structured output.
Moonshot AIKimi K2.5
Balanced general-purpose model from Moonshot AI. Solid tool use and instruction following at a competitive cost.
Moonshot AIKimi K2 Thinking
Moonshot's reasoning-focused model with extended chain-of-thought. Best for complex analysis that benefits from visible step-by-step reasoning.
OpenAIGPT-4o Mini
OpenAI's budget-friendly model. Fast and capable for straightforward tasks like email drafts, data formatting, and simple lookups.
OpenAIGPT-5.3
OpenAI's code-optimized model. Generates, debugs, and refactors code with high accuracy across multiple languages and frameworks.
OpenAIGPT-5.4
OpenAI's well-rounded flagship. Strong at reasoning, tool use, and structured output — a reliable default for agent workflows.
OpenAIGPT-5.4 Mini
Scaled-down GPT-5.4 that trades some capability for significantly lower cost and faster response times.
OpenAIGPT-5.4 Nano
OpenAI's smallest and cheapest model. Ideal for lightweight, high-volume tasks like classification and simple extraction.
OpenAIGPT-5.4 Pro
Extended compute variant of GPT-5.4. Allocates more reasoning budget for the most complex analytical and coding tasks.
OpenAIGPT-5.5
Next-gen OpenAI model with stronger reasoning and better instruction following than GPT-5.4. Handles complex multi-step agent tasks.
OpenAIGPT-5.5 Pro
Extended compute GPT-5.5 for the most demanding tasks. Premium pricing for premium performance on hard reasoning problems.
OpenAIGPT-5.6 Luna
OpenAI's efficiency-optimized variant. Fast, affordable, and capable — designed for high-throughput automation at scale.
OpenAIGPT-5.6 Luna Pro
Enhanced Luna with more reasoning depth. A middle ground between Luna's speed and Sol's capability.
OpenAIGPT-5.6 Sol
OpenAI's balanced workhorse model. Strong all-around performance for code, analysis, and multi-step agent workflows.
OpenAIGPT-5.6 Sol Pro
Extended compute Sol for heavier reasoning tasks. When Sol isn't enough but you don't need Terra.
OpenAIGPT-5.6 Terra
OpenAI's top-tier reasoning model. Maximum capability for the hardest analytical, coding, and planning tasks.
OpenAIGPT-5.6 Terra Pro
Full compute GPT-5.6. The most capable OpenAI model available — reserved for the most complex agent workflows.
xAIGrok 4.5
xAI's flagship model with real-time web knowledge. Strong at current events, research, and tasks that benefit from live context.
XiaomiMiMo-V2.5
Xiaomi's efficient reasoning model. Competitive performance at very low cost, optimized for structured output tasks.
XiaomiMiMo-V2.5 Pro
Enhanced MiMo variant with improved reasoning depth and instruction following for more complex agent workflows.
Zhipu AIGLM 5.2
Zhipu AI's general-purpose model with strong multilingual capabilities. A solid choice for tasks requiring Chinese and English fluency.