Explore Text Models

DeepSeek V4 Pro Cheaper is the same large-scale Mixture-of-Experts model from DeepSeek with a 1M-token context window, built for advanced reasoning, coding, long-horizon agent workflows, knowledge, math, and software engineering tasks. This variant is temporarily cheaper while DeepSeek has lowered its prices, and can route via Chinese providers, so data privacy is not guaranteed.

DeepSeek V4 Pro Cheaper Thinking enables DeepSeek's chain-of-thought mode on the same large-scale Mixture-of-Experts model with a 1M-token context window, built for advanced reasoning, coding, long-horizon agent workflows, knowledge, math, and software engineering tasks. This variant is temporarily cheaper while DeepSeek has lowered its prices, and can route via Chinese providers, so data privacy is not guaranteed.

DeepSeek V4 Pro in a Trusted Execution Environment (TEE) with attestation support. Built for advanced reasoning, coding, long-horizon agent workflows, knowledge, math, and software engineering tasks.

DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with a 1M-token context window, built for fast inference, high-throughput workloads, reasoning, coding, and agent workflows.

DeepSeek V4 Flash Thinking enables DeepSeek's reasoning mode on the efficiency-optimized Mixture-of-Experts model with a 1M-token context window, built for fast inference, high-throughput workloads, reasoning, coding, and agent workflows.

DeepSeek V4 Pro is a large-scale Mixture-of-Experts model from DeepSeek with a 1M-token context window, built for advanced reasoning, coding, long-horizon agent workflows, knowledge, math, and software engineering tasks.

DeepSeek V4 Pro Thinking enables DeepSeek's chain-of-thought mode on the large-scale Mixture-of-Experts model with a 1M-token context window, built for advanced reasoning, coding, long-horizon agent workflows, knowledge, math, and software engineering tasks.

GPT-5.5 is OpenAI's smartest and most intuitive model yet, built for agentic coding, computer use, and professional knowledge work with stronger reasoning and token efficiency.

Qwen3.6 27B is a native vision-language dense model with stronger agentic coding and STEM reasoning than Qwen 3.5 27B. It also improves spatial intelligence (including object localization/detection), plus video understanding, document OCR, and visual-agent workflows.

Ling-2.6-1T is an inclusionAI instruction model optimized for large-scale agentic and coding workloads with long-context support and structured output capabilities.

Hy3 Preview is a high-efficiency Mixture-of-Experts model from Tencent designed for agentic workflows and production use. It supports configurable reasoning levels to balance speed and depth, with strong code generation performance in multi-step workflows.

Open source preview release in Arcee's Trinity Large family with a 262K context window, strong tool use, and production-grade latency for agent workflows.

MiMo V2.5 is Xiaomi's full-modal understanding model for agent workflows. It supports deep reasoning, tool calling, structured outputs, and web search with up to 1M context.

MiMo V2.5 Pro is Xiaomi's long-context flagship general model for coding and agentic orchestration. It supports reasoning, tool calling, and structured outputs with up to 1M context.

Qwen3 Coder Next is optimized for coding agents and local development workflows with a sparse 80B MoE architecture. Running inside a TEE (Trusted Execution Environment), with provider attestation support.

Kimi K2.6 is Moonshot AI's next-generation multimodal model designed for long-horizon coding and agentic orchestration. Running inside a TEE (Trusted Execution Environment), with provider attestation support.

Ling-2.6-flash is an instant (instruct) model from inclusionAI with 104B total parameters and 7.4B active parameters, designed for real-world agents that require fast responses, strong execution, and high token efficiency. It delivers performance comparable to state-of-the-art models at a similar scale while significantly reducing token usage across coding, document processing, and lightweight agent workflows.

Qwen3.6 Max Preview is Alibaba's flagship Qwen 3.6 model for complex tasks. It supports both thinking and non-thinking modes in a single model id.

MiniMax M2.5 is a productivity-focused flagship model with stronger coding and office workflow performance (Word, Excel, PowerPoint), plus improved tool planning and token efficiency. Running inside a TEE (Trusted Execution Environment), with provider attestation support.

GLM-5.1 is Z.AI's next-gen model optimized for long-horizon agent workflows and deep reasoning. Running inside a TEE (Trusted Execution Environment), with provider attestation support.

GLM-5.1 with extended reasoning mode for multi-step planning and tool-heavy workflows. Running inside a TEE (Trusted Execution Environment), with provider attestation support.

Xiaomi MiMo V2 Flash (309B MoE, 15B active) served with TEE-backed inference and attestation support.

Qwen3.6 35B A3B is a native vision-language MoE model with hybrid attention. Compared to Qwen3.5 35B A3B, Alibaba reports stronger agentic coding, mathematical and code reasoning, and better spatial understanding (including object localization and detection).

Qwen3.6 35B A3B is a native vision-language MoE model with hybrid attention. Compared to Qwen3.5 35B A3B, Alibaba reports stronger agentic coding, mathematical and code reasoning, and better spatial understanding (including object localization and detection).

Qwen3.6 Flash is Alibaba's fast native vision-language model in the Qwen 3.6 family. It improves over 3.5 Flash with stronger coding/agent performance and better spatial intelligence, including object localization and detection.

Kimi K2.6 is an open-source, native multimodal agentic model built for long-horizon coding, coding-driven design, and large-scale task orchestration. It can turn simple prompts and visual inputs into production-ready interfaces and full-stack workflows, and is designed to coordinate complex multi-agent plans with thousands of steps across code, documents, and spreadsheets.

Kimi K2.6 Thinking is the reasoning-optimized K2.6 variant for deeper multi-step planning and execution. It is tuned for long-horizon coding and design workflows, including complex orchestration across many specialized sub-agents and autonomous end-to-end output generation.

Claude Opus 4.7 is a major upgrade for advanced software engineering, long-running complex tasks, and high-resolution vision understanding.

Claude Opus 4.7 with thinking enabled (default budget: 16k tokens).

Step 3.5 Flash 2603 is optimized for high-frequency agentic and coding workflows with improved token efficiency and faster reasoning. NOTE: This model runs via StepFun, which may log and train on your prompts.