一个网关，统一所有 模型。

SpaceXAI's frontier model for coding, agentic tasks, and knowledge work — succeeds Grok 4.6 with mandatory adjustable reasoning (low/medium/high/xhigh, default high).

TypeSafe Jev is a System One decision model: send application state and typed questions (noul / choice / score) and get structured answers with probabilities — not generated chat text. Call POST /api/v1/systemone. Requires a TypeSafe BYOK key (dashboard → BYOK → TypeSafe). List price is $0.042 per million input tokens; output tokens are free upstream. This hop is user-key only (0/0 on AnyRouter credits).

Union Alpha is a stealth preview multimodal model for research, coding, and agentic workflows. It offers frontier-level general-purpose performance with vision in and text out. OpenRouter lists tools and structured output; it does not advertise reasoning on this SKU.

Agnes 3.0 Flash is designed for real-world agent tasks and development workflows, covering the full execution chain from task understanding and planning to tool invocation and final delivery. The model focuses on improving stability, instruction-following, factual grounding, and output completeness in complex tasks, with vision in and text out via an OpenAI-compatible Chat Completions interface.

NVIDIA Nemotron Parse 2.0 is a document-understanding VLM: page image in, structured text out (markdown, layout classes, bounding boxes, reading order). Served via NVIDIA NIM OpenAI-compatible /v1/chat/completions. Not image generation.

Ling-3.0-flash-Fin is inclusionAI's finance-focused mixture-of-experts model, built on Ling-3.0-flash (124B total / 5.1B active). It is designed for real-world investment research, market analysis, and financial reasoning, while retaining general reasoning, coding, and agentic skills. Distinct from text-only inclusionai/ling-3.0-flash, the VL listing, and the health/medicine Sante SKU. Served free via the platform free pool, with BYOK as a fallback.

Ling-3.0-flash-Sante is inclusionAI's health and medicine-focused mixture-of-experts model, built on Ling-3.0-flash (124B total / 5.1B active). It is designed for medical knowledge reasoning, clinical safety, evidence-based retrieval, and long-horizon medical tasks, while retaining general reasoning, coding, and agentic skills. Distinct from text-only inclusionai/ling-3.0-flash and the VL listing. Served free via the platform free pool, with BYOK as a fallback. Command Code lists the same product as a free-while-it-lasts promo — that is the upstream's credit, not AnyRouter Free monthly credits.

Ling-3.0-flash-VL is inclusionAI's native multimodal instruct model, a 124B-parameter Mixture-of-Experts model with roughly 5.5B activated parameters per token. Built on Ling-3.0-flash, it adds native image and video understanding (up to 256K context) for visual reasoning, document and chart reading, and agentic GUI tasks. Distinct from the text-only inclusionai/ling-3.0-flash listing. Served free via the platform free pool, with BYOK as a fallback.

Nex-N2.5-Mini is Nex AGI's smaller agentic coding model (35B total / 3B active MoE), built for high-speed instruction following, real-time tool execution, and visually assisted computer use. Distinct from Nex-N2.5-Pro and the retired Nex-N2-Pro listing. OpenRouter lists a :free SKU only — that wire is an alias, not the catalog id. Served free via the platform free pool, with BYOK as a fallback.

Nex-N2.5-Pro is Nex AGI's larger agentic coding model (397B total / 17B active MoE), built to turn goals into working, verified outcomes with a visual feedback loop. Distinct from Nex-N2.5-Mini and the retired Nex-N2-Pro listing. OpenRouter lists a :free SKU only — that wire is an alias, not the catalog id. Served free via the platform free pool, with BYOK as a fallback.

DeepSeek V4.1 Flash is the cost-efficient sparse MoE tier of the V4.1 family (1M context, vision in, text out). Official API id is `deepseek-flash`; OpenRouter lists the same product as `deepseek/deepseek-v4.1-flash`.

OpenAI GPT-6 Astra reasoning/chat model with ~1.05M context; Experiential Cloud promotional free daily tier.

Meta's Muse Spark 1.3 is a multimodal reasoning model for complex agentic tasks. It accepts text, images, video, audio, and PDF documents, returns text, and offers a 1M-token context window.

Google's Gemini 3.8 Flash is a multimodal model for fast agentic workflows, coding, and complex multi-step reasoning, with a 1,048,576-token context window.

Inception Labs Mercury 2.5 Preview is a text model with thinking and tool use and a 260,000-token context window.

Anthropic's Claude Fable 5.1 is the successor to Claude Fable 5 for long-running agentic coding, knowledge work, and research. Same $10/$50 list rates as Fable 5, with cache reads at $0.25 per million tokens and a 1M-token context window.

Alibaba Qwen native vision-language model for coding, office, long-context reasoning, and agents. 1M context.

Hy4 Preview is Tencent Hunyuan's 770B/49B-active MoE for agents, coding, office automation, and complex tool use, with a 1M-token context window.

Agnes 2.5 Flash is Agnes AI's fast and efficient language model, optimized for coding tasks, agent workflows, tool calls, multi-turn dialogue, reasoning, and image understanding via an OpenAI-compatible Chat Completions interface.

Agnes 2.5 Pro Alpha is Agnes AI's paid inference model for advanced coding, scientific reasoning, long-context analysis, agent workflows, and multimodal understanding via an OpenAI-compatible Chat Completions API.

Agnes 2.5 Pro is Agnes AI's paid inference model and the commercially stable version of the Agnes 2.5 Pro Alpha ranking model, suited for advanced coding, scientific reasoning, long-context analysis, agent workflows, and multimodal understanding.

MAI-Thinking-1 is Microsoft's first inference model in the MAI series, built for enterprise-scale workloads with strong reasoning, mathematical, and general intelligence capabilities at high throughput.

GLM-5.3-Flash is a native multimodal model from Z.ai (320B-A18B, MIT License, 1M-token context). It is suited for efficient coding and long-horizon agent tasks. Its hybrid sparse and linear attention architecture maintains accurate long-context behavior while reducing compute overhead.

GLM-5.3 is Z.AI's reasoning model for coding and agentic workflows, built for complex software engineering, long-running agents, and vulnerability analysis. It uses the same base model as GLM-5.2 with scaled post-training for stronger coding performance, task execution, and token efficiency.