AI Models
DeepSeek
Chinese lab shipping open-weight MoE models. The V4 family (previewed April 24, 2026) runs a 1M context on DeepSeek's API; R1 (2025) popularized open reasoning models. Low-cost API.
Try DeepSeekVisit deepseek.comdeepseek.com
Share
DeepSeek produces open-weight language models that punch far above their weight class. DeepSeek-R1, released in 2025, showed an open-weight reasoning model could compete with OpenAI's o1 on math, code, and science benchmarks at a fraction of the cost. The current API models are DeepSeek-V4.1-Flash (released September 10, 2026 and served as deepseek-flash) and DeepSeek-V4-Pro (GA August 13, 2026), both with a 1M token context window; the original V4-Flash from the April 24, 2026 preview has been retired. The models use mixture-of-experts architecture, which keeps inference costs low despite their large total parameter counts. You can run them locally via Ollama, access them through their own API, or use them on OpenRouter. For developers building cost-sensitive AI applications or teams that need to self-host, DeepSeek models are consistently among the strongest performance-per-dollar options.
DeepSeek at a glance
Last verified
- Pricing
- deepseek-flash (V4.1-Flash) - $0.30/1M input (cache miss), $0.006 cache hit, $1.20/1M output at peak; off-peak is half price
- deepseek-v4-pro - $1.32/1M input (cache miss), $0.044 cache hit, $3.96/1M output at peak; off-peak is half price
- Peak hours are 01:00-04:00 and 06:00-10:00 UTC, Monday to Friday, excluding Chinese public holidays; all other hours, including weekends and Chinese public holidays, bill at off-peak rates
- Install
pip3 install openainpm install openai- Models
- The API serves deepseek-flash (DeepSeek-V4.1-Flash-1017, with vision) and deepseek-v4-pro, both with a 1M context window and up to 384K output tokens, through OpenAI-format and Anthropic-format endpoints. Thinking mode is on by default, with low, high and max effort levels; the legacy deepseek-v4-flash names now route to V4.1-Flash.
- Recent changes
- DeepSeek-V4.1-Flash-1017 update ships to deepseek-flash and the web and app, with better agentic coding scores and shorter outputs at unchanged prices.
- DeepSeek-V4.1-Flash releases: a 552B-parameter MoE with native vision and open weights on Hugging Face. It retires V4-Flash and V4-Flash-Vision-Exp and cuts API prices.
- DeepSeek-V4-Pro reaches GA with native Responses API support for Codex and low, high and max thinking effort levels; peak/off-peak pricing starts August 16.
Sources: DeepSeek Models & Pricing, DeepSeek Your First API Call, DeepSeek API change log, DeepSeek-V4.1-Flash announcement
Similar Tools
AI Models
Claude
Anthropic's model family. As of October 2026: Opus 5.5 as the default, Fable 5.1 for the hardest long-horizon work, Sonnet 5.5 for speed, Haiku 5.5 for cost. 1M context. Max from $100/mo.
AI Models
G
GPT-5
OpenAI's GPT-5 model, released August 2025. Unified fast and reasoning modes in one model. Superseded by the GPT-6 family in September 2026 and now marked deprecated in the API.
AI Models
Gemini
Google's model family. As of October 2026 Gemini 4 Argon (GA October 13, 1M token context) is the flagship for coding and agents, with Gemini 3.8 Flash (free tier in AI Studio) as the fast, cheaper option.
AI Models
Grok
xAI's model family, now documented under the SpaceXSI (SpaceX Super Intelligence) name. Grok 4.7 (September 21, 2026) has a 500K context at $2/$6 per million tokens. OpenAI-compatible API.
Get started with DeepSeek
Chinese lab shipping open-weight MoE models. The V4 family (previewed April 24, 2026) runs a 1M context on DeepSeek's API; R1 (2025) popularized open reasoning models. Low-cost API.
Try DeepSeekGet weekly tool reviews
Honest takes on AI dev tools, frameworks, and infrastructure - delivered to your inbox.
Subscribe FreeMore AI Models Tools
Claude
Anthropic's model family. As of October 2026: Opus 5.5 as the default, Fable 5.1 for the hardest long-horizon work, Sonnet 5.5 for speed, Haiku 5.5 for cost. 1M context. Max from $100/mo.
ChatGPT
OpenAI's chat app. As of October 2026 Free and Go run GPT-6 Luna; Plus ($20/mo) adds GPT-6.1 Sol and GPT-6 Astra. Agent mode, Codex, image generation, code interpreter.
OpenRouter
Unified API for hundreds of models. One API key, one billing dashboard. OpenAI, Anthropic, Google, Meta, Mistral, and more. Automatic fallbacks and load balancing.
Related Guides
Guide
Run AI Models Locally with Ollama and LM Studio
Install Ollama and LM Studio, pull your first model, and run AI locally for coding, chat, and automation - with zero cloud dependency.
Getting StartedGuide
Model Aliases - Claude Code
Use opus, sonnet, haiku, and best to switch models easily.
Claude CodeGuide
Getting Started with DevDigest CLI
Install the dd CLI and scaffold your first AI-powered app in under a minute: Next.js 16, Convex, Clerk, Autumn, Tailwind v4 and an agent-ready CLAUDE.md.
Getting StartedRelated Posts
7 min read
News
Ornith-1.0: What an Open Source Self-Improving Coding Model Actually Means
Ornith-1.0 open-source coding models explained: what "self-improving" really means, the four model sizes, hardware needs...
8 min read
ai
DiffusionGemma: Google Bets Diffusion Can Make Text Generation 4x Faster
DiffusionGemma is Google's 26B MoE open model that generates 256-token blocks in parallel. What diffusion generation mea...
8 min read
Qwen
Qwen3.6-27B Is the Local Coding Model to Test First
Qwen3.6-27B is the local coding model to benchmark first: what hardware it needs, which jobs it should take over, and wh...
8 min read
OpenAI
OpenAI's GPT 5.4 in 10 Minutes
State-of-the-art computer use, steerable thinking you can redirect mid-response, and a million tokens of context. GPT 5....
4 min read
Cursor
Cursor 2.0 & Composer: The Fastest AI Coding Model
Cursor 2.0 ships Composer, the first in-house Cursor coding model, and a multi-agent interface. How Composer was trained...
7 min read
Kimi
Kimi K2: Fast, Cheap, and Efficient Coding
Kimi K2 from Moonshot AI replaced Claude Sonnet 4 in my stack - an open-source 1T-parameter MoE model that is faster, ch...
Try the member tools free
Get Smarter About AI Dev
New tutorials, open-source projects, and deep dives on coding agents - delivered weekly.
One email per weekReal code, not theoryFree forever