Loading Tavory AI
DeepSeek model family
Compare DeepSeek models represented in Tavory for coding, reasoning and cost-aware chat.
This hub organizes distinct DeepSeek versions into a focused comparison instead of exposing repeated provider routes as separate search pages.
Tavory is an independent AI workspace. DeepSeek models are developed by DeepSeek; Tavory provides catalog comparison and access to eligible routes.
Explore DeepSeek in TavoryCurrent Tavory catalog
Each card links to a canonical model page with current catalog context, capabilities and price signals. Regional duplicates stay available to Tavory's router but do not create repetitive public pages.
DeepSeek
An alias for DeepSeek v4 Flash in non-thinking mode.
DeepSeek
Strong open reasoning model
DeepSeek
An alias for DeepSeek v4 Flash in thinking mode.
Deepinfra
DeepSeek-V3, a strong Mixture-of-Experts (MoE) language model with 671B total parameters with 37B activated for each token. To achieve efficient inference and cost-effective training, DeepSeek-V3 adopts Multi-head Latent Attention (MLA) and DeepSeekMoE architectures.
DeepSeek
Affordable and versatile assistant
Vertex
DeepSeek-V3.2 is a model that harmonizes high computational efficiency with superior reasoning and agent performance. DeepSeek's approach is built upon three key technical breakthroughs: DeepSeek Sparse Attention (DSA), scalable reinforcement learning framework, and large scale agentic task synthesis pipeline.
DeepSeek
DeepSeek v4 Flash - 284B total / 13B active params. Your fast, efficient, and economical choice.
Tensorx
DeepSeek V4 Flash 0731 is the July 31 update of DeepSeek's efficiency-optimized MoE model (284B total, 13B active parameters) with a 1M-token context window. It is built for fast inference and high-throughput workloads while keeping strong reasoning and coding performance. Reasoning is toggleable and off by default.
Doubleword
DeepSeek V4-Flash is the smaller model in the V4 family, ahead of models like Gemini 3 Flash and Qwen3.5-397B-A17B.
Tensorx
DeepSeek V4 Pro is a large scale Mixture of Experts model from DeepSeek with 1.6T total parameters and 49B activated parameters, supporting a 1M token context window. It is built for advanced reasoning, coding, and long horizon agent workflows, with strong performance across knowledge, math, and software engineering benchmarks. Built on the same architecture as DeepSeek V4 Flash, it adds a hybrid attention system for efficient long context processing. Reasoning efforts high and xhigh are supported, with xhigh mapping to max reasoning. It is well suited for complex workloads such as full codebase analysis, multi step automation, and large scale information synthesis, where both capability and efficiency are critical.
Doubleword
DeepSeek V4-Pro is DeepSeek's flagship open-weights MoE model, with performance in the same class as Claude Sonnet 4.6 and stronger overall results than GLM-5.1.
Check whether the task needs files, images, tools, reasoning or a specific output type.
Use context and provider price signals as constraints, not as guarantees of answer quality.
Tavory checks route health, plan access and final cost again before the request runs.
