Models

GLM-5.3 is Z.ai’s reasoning model for coding and agentic workflows, designed for complex software engineering, long-running agents, and vulnerability analysis. It uses the same base model as GLM-5.2, with scaled post-training improving coding, task execution, and token efficiency. This model is a limited-time preview version of GLM-5.3, intended for testing and evaluation only. Service stability is not guaranteed, and we do not recommend using it in production environments. We’re waiting for the official commercial API release and will integrate it as soon as official support becomes available.

Gemini 3.7 Flash is Google’s natively multimodal reasoning model for coding, agents, web development, and knowledge work. It supports a 1M-token context window and adjustable thinking levels. Compared with Gemini 3.6 Flash, it improves coding, tool use, multi-step planning, and instruction following.

Gemini 3.7 Flash free version: fFree model resources are limited and provided only for trial use; stability cannot be guaranteed, and you may encounter 429 errors during use. If you need to use it in a production environment and require unlimited concurrency with absolute stability, please choose the official version: gemini-3.7-flash

GLM-5.2 is Z.ai’s flagship model for the era of long-horizon tasks. With a truly usable 1M-token context window, it can handle project-level engineering context, execute long-running tasks more reliably, follow engineering standards more consistently, and complete the full development workflow from requirements to multi-platform deployment in a single task.

DeepSeek-V4-Flash-0731(deepseek-v4-flash-0731) is an open-source MoE large language model developed by the Chinese AI company DeepSeek, with support for a million-token context window. It is designed for coding, complex reasoning, tool use, agentic workflows, and long-document processing. Its advantages include strong performance with fewer active parameters and improved efficiency through DSpark speculative decoding. Compared with DeepSeek V4-Flash Preview, it offers significantly stronger coding and agent capabilities, while outperforming DeepSeek V4-Pro Preview on several benchmarks with fewer active parameters.

DeepSeek V4 Pro 0813 is DeepSeek’s high-performance general-purpose reasoning and agent model, designed for complex reasoning, coding, long-document analysis, and agentic workflows. It supports thinking and non-thinking modes, a 1M-token context window, up to 384K output, tool calling, and the Responses API. Compared with V4 Flash 0731, Pro prioritizes capability on complex tasks, while Flash focuses on speed, cost efficiency, and high concurrency.

Grok 4.6 is xAI’s (SpaceXAI) flagship multimodal reasoning model for coding, long-running agents, knowledge work, and interactive application development. It supports image understanding, a 500K context window, tool calling, and structured outputs. Compared with Grok 4.5, it offers stronger multi-step execution, self-verification, coding, and visual project generation.

Seedance 2.5 is ByteDance Seed’s next-generation unified multimodal audio-video generation model, designed for filmmaking, advertising, education, simulation, and long-form content creation. It generates up to 30 seconds of synchronized audio and video in one pass and supports multi-round extensions. Compared with Seedance 2.0, it offers stronger storytelling, smoother transitions, richer reference support, improved realism, and more precise editing.

GPT-5.6 Luna is designed for cost-sensitive, high-volume workloads. It roughly corresponds to the nano model tier used in earlier GPT-5 families.

GPT‑5.6 Sol sets a new standard for both intelligence and efficiency, achieving state-of-the-art results across coding, knowledge work, cybersecurity, and science while outperforming previous and competing frontier models with fewer tokens and at lower estimated cost.

GPT-5.6 Terra is designed for workloads that balance intelligence and cost. It roughly corresponds to the mini model tier used in earlier GPT-5 families.

Agnes 2.5 Flash is Agnes AI’s fast and efficient language model, an upgraded, fully available model based on Agnes 2.0 Flash. It continues to use an OpenAI-compatible Chat Completions interface and has been optimized for coding tasks, agent workflows, tool calls, multi-turn dialogue, reasoning, and image understanding.

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, suitable for advanced coding, scientific reasoning, long-context analysis, agent workflows, and multimodal understanding. The model is accessed via an OpenAI-compatible Chat Completions API.

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

Agnes Image 2.1 Flash is Agnes AI’s high-performance image generation and image editing model, supporting text-to-image, image-to-image, and multi-image composition. It is suitable for creative design, marketing visuals, e-commerce product images, and social content production.

Grok 4.5 was trained on datasets spanning knowledge in coding, science, engineering, and math. With both intelligent and efficient reasoning, Grok 4.5 excels at real engineering tasks and exceeds comparable leading models at these tasks.

LFM-2.5-2.6B is a compact reasoning model developed by Liquid, featuring a generous 128,000 token context length. It is highly suited for agent workflows, data extraction, retrieval-augmented generation (RAG), and long-context processing. However, the developer advises against using this model for agentic coding tasks.

MiniMax H3 (minimax-h3) is a general-purpose omni-modal generation model developed by the Chinese AI company MiniMax. It can jointly understand text, images, video, and audio while generating videos of up to 15 seconds in 2K resolution with native stereo sound. It is designed for advertising, branding, e-commerce, product design, UI/UX, and gaming. Compared with Hailuo 01 and Hailuo 02, H3 evolves from specialized video generation into a unified model for multimodal creation, reference-based editing, text rendering, and motion transfer.

Muse Glimmer 30B is a dense causal language model distilled from Muse Spark, designed for autonomous agent work. It combines multi-step reasoning, robust pattern-based tool invocation, and fault-recovery capabilities, and offers multimodal understanding via a ViT-G/14 perceptual encoder of roughly 1.8 billion parameters, supporting interleaved text and image inputs, a context window of over 131K tokens, and optional inference strength settings (from low to ultra-high). Muse Glimmer is trained on data in over 100 languages, performs strongly for its size on agent benchmarks including MCP Atlas, DeepSearch QA, Gaia2, and SWE-Bench Pro, and is released under the Apache 2.0 license.