Model Directory

Aymo includes access to more than forty leading AI models. Each model offers unique strengths across reasoning, coding, writing, analysis, or high-speed tasks.

This directory lists every model available in Aymo and what each model is best suited for.

GPT Models

High-accuracy models designed for reasoning, writing, analysis, and advanced language tasks.

• GPT-5: Great for complex reasoning, long-form writing, technical analysis, and high-accuracy outputs.
• O3: Great for structured reasoning, math, logic, and step-by-step problem solving.
• GPT-5 Mini: Great for balanced reasoning with lower cost and fast performance.
• GPT-5 Nano: Great for quick replies, lightweight tasks, and high-volume usage.
• O4 Mini: Great for compact reasoning tasks with strong speed.
• O3 Mini: Great for lightweight logic tasks and simple structured outputs.
• GPT OSS 120B: Great for open-source workflows, transparent behavior, and general reasoning.
• GPT-4.1 Mini: Great for fast writing, summarization, and efficient analysis.
• GPT-4.1 Nano: Great for everyday tasks with minimal latency.
• GPT-4o Mini: Great for fast multimodal tasks with small image-text inputs.
• GPT-3.5 Turbo: Great for general tasks, summaries, and basic coding.

Claude Models

Models focused on safe outputs, high-quality writing, and consistent reasoning.

• Claude Sonnet 4.5: Great for advanced writing, structured reasoning, and factual stability.
• Claude Haiku 4.5: Great for fast outputs with strong clarity.
• Claude 3.7 Sonnet: Great for refined logic, long context workflows, and stable performance.
• Claude 3.5 Haiku: Great for everyday tasks and high-speed writing.

Gemini Models

Multimodal models optimized for reasoning, coding, and vision-based tasks.

• Gemini 2.5 Pro Preview: Great for multimodal reasoning, coding, and advanced analysis.
• Gemini 2.5 Pro: Great for long-context tasks, visual understanding, and complex reasoning.
• Gemini 2.5 Flash: Great for fast responses and cost-efficient writing.
• Gemini 2.0 Flash: Great for stable summaries and structured outputs.
• Gemini 2.5 Flash Lite: Great for ultra-fast, low-latency tasks.
• Gemma 3 1B Instruct: Great for simple writing and lightweight tasks.

Grok Models

Models focused on speed, reasoning, and developer-friendly outputs.

• Grok 4 Fast: Great for rapid reasoning and quick practical tasks.
• Grok 3: Great for direct, simple conversational answers.
• Grok Code Fast 1: Great for code generation and debugging.
• Grok 3 Mini: Great for lightweight reasoning and small coding tasks.

Qwen Models

Large reasoning, coding, and vision-language models with strong multilingual ability.

• Qwen3 Max: Great for high-accuracy reasoning and multilingual tasks.
• Qwen3 Coder 480B A35B Instruct: Great for enterprise-level code generation.
• Qwen3 VL 235B A22B Thinking: Great for visual reasoning and multimodal analysis.
• Qwen3 VL 235B A22B Instruct: Great for instruction-focused multimodal tasks.
• Qwen3 235B A22B Thinking 2507: Great for deep reasoning with long context.
• Qwen3 235B A22B Instruct 2507: Great for structured instruction tasks.
• Qwen3 Next 80B A3B Thinking: Great for mid-scale long-form reasoning.
• Qwen3 Next 80B A3B Instruct: Great for writing, coding, and structured outputs.
• Qwen3 Coder 30B A3B Instruct: Great for fast mid-scale code generation.
• Qwen 3 4B FP8: Great for low-cost, ultra-light tasks.

Kimi Models

Reasoning-focused and instruction-tuned models with stable performance.

• Kimi K2 Thinking: Great for analytical tasks and deep reasoning.
• Kimi K2 Instruct: Great for instruction-following and productivity tasks.
• Kimi K2 0905: Great for predictable, stable text generation.

Llama Models

Lightweight open models for writing, coding, and multilingual tasks.

• Llama 4 Scout 17B: Great for efficient multilingual work and mid-scale reasoning.
• Llama 3.1 8B Instruct: Great for lightweight writing and quick code output.

Mistral Models

Small, efficient models for fast generation and general tasks.

• Mistral Nemo: Great for fast writing, summaries, and efficient content generation.

Perplexity Models

Models designed for retrieval-enhanced responses and up-to-date answers.

• Perplexity Sonar: Great for grounded results, research tasks, and retrieval-based answers.

DeepSeek Models

Models focused on experimental reasoning and high-detail analysis.

• DeepSeek V3.2 Exp: Great for structured analysis, technical insights, and reasoning tasks.

MiniMax Models

Compact bilingual models optimized for efficient generation.

• MiniMax M2: Great for English-Chinese tasks and efficient daily writing.

Ernie Models

Knowledge-driven models for structured information tasks and multilingual QA.

• Ernie 4.5 21B: Great for reasoning, multilingual work, and knowledge-based outputs.

Together, these models create a complete multi-model environment inside Aymo. The platform makes it possible to switch models instantly and apply the most effective option for any task or workflow.