Model Library
Explore the live Qwen3 model and the open-model roadmap for idyl inference.
Live now. A fast general-purpose Qwen3 model for chat, coding, multilingual prompts, and reasoning/non-reasoning workflows through the OpenAI-compatible API.
Roadmap family for long-context coding agents, repository-scale prompts, tool use, and efficient MoE serving.
Roadmap reasoning family for math, code, planning, and distillation-friendly workloads where explicit step-by-step reasoning is valuable.
Roadmap general-purpose MoE family for strong coding, multilingual chat, structured generation, and cost-efficient high-capacity serving.
Roadmap multimodal Llama family for text and image understanding, broad language coverage, and long-context application workflows.
Roadmap family spanning efficient chat, multimodal understanding, specialist reasoning, and Devstral-style software agents.
Roadmap agentic family for tool-heavy reasoning, coding, visual workflows, and large-context automation.
Roadmap family for efficient text and image applications, low-latency chat, summarization, and multimodal question answering.
Roadmap compact reasoning model for math, science, coding, and latency-sensitive applications that still need deliberate reasoning behavior.
No models found
Try a different search term or filter.