DeepSeek
Verified Oct 16, 2026 ActiveOpen-weight reasoning models with a free chat app
Listed on EffectHub ·
Pricing
- Pricing model
- Freemium
The chat app is free (App Store listing shows Free). The API is prepaid per token with no monthly plan: deepseek-flash costs $0.30 per million input tokens on a cache miss ($0.006 on a cache hit) and $1.20 per million output at peak, half that off-peak; deepseek-v4-pro costs $1.32 input and $3.96 output per million at peak. Peak hours are 01:00-04:00 and 06:00-10:00 UTC on Chinese working days.
Use tool ↗About DeepSeek
DeepSeek is a Hangzhou research lab that publishes open-weight language and reasoning models and serves them three ways: a free chat assistant on the web and in iOS and Android apps, a desktop client called Harness, and a pay-per-token API on its open platform. The current line is DeepSeek-V4.1-Flash, a fast model with native image understanding, and the larger DeepSeek-V4-Pro, both with a 1M-token context window; earlier V3 and R1 releases drew attention for matching much larger models at a fraction of the training cost. API prices are among the lowest of any frontier lab and halve again in off-peak hours, which is why it shows up as the default cheap backend in many agent and coding tools. The service runs under the law of mainland China.
- API
- Yes
- Last verified
- Oct 16, 2026
Verified sources
- API pricing page: deepseek-flash $0.15/$0.30 per million input tokens (cache miss, off-peak/peak), $0.003/$0.006 on cache hit, $0.60/$1.20 output; deepseek-v4-pro $0.66/$1.32 input and $1.98/$3.96 output; 1M context; peak hours 01:00-04:00 and 06:00-10:00 UTC (15 October 2026)api-docs.deepseek.com ↗
- Home page: web chat, mobile app, Harness desktop client and API platform; DeepSeek-V4.1-Flash newest with native visual understanding; models V4, V3.2, V3.1, R1, V3; copyright Hangzhou DeepSeek Artificial Intelligence Basic Technology Research Co., Ltd. (15 October 2026)deepseek.com ↗
- App Store listing: DeepSeek - AI Assistant is Free, seller Hangzhou DeepSeek Artificial Intelligence Co., Ltd. (15 October 2026)apps.apple.com ↗
- Terms of use: operated by Hangzhou DeepSeek Artificial Intelligence Co., Ltd., governed by the laws of the People's Republic of China in the mainland (15 October 2026)cdn.deepseek.com ↗
- High-Flyer announced the establishment of DeepSeek in July 2023lex.dk ↗
Is DeepSeek legit?
We don’t label tools safe or unsafe. Below are the records we checked ourselves, each with the date we read it.
- Domain registered
- Aug 13, 2000 (26 years ago)Registry record (RDAP), registrar DNSPod, Inc., looked up Sep 6, 2026
- TLS certificate
- Issued by Amazon, valid until Dec 25, 2026Seen in our own connection to the site on Oct 16, 2026
- Last checked
- Oct 16, 2026The site answered
- On the market since
- 2023The vendor's own material, read by us
- Operated by
- Hangzhou DeepSeek Artificial Intelligence Basic Technology Research Co., Ltd., ChinaOur research, card verified Oct 15, 2026
Pros & cons of DeepSeek
Pick DeepSeek for the price-performance of the API and the open weights, not for the chat app, which is a plain assistant without the connectors and workspace features the big consumer products now carry. For a developer paying per token it is hard to beat; for a company with data-residency rules the China-hosted service and PRC-law terms settle the question before quality does.
Pros
- Flash output at $1.20 per million tokens peak and $0.60 off-peak undercuts every Western frontier API by a wide margin
- Open-weight releases mean the models can be run locally or on any cloud, with no lock-in to the hosted service
- 1M-token context on both current models, with cached-input pricing that makes long agent loops cheap
- Chat app is free on web, iOS and Android with no tier to upgrade to
Cons
- Terms are governed by the law of mainland China and the operator is a Hangzhou company, which rules it out for many regulated buyers
- Peak and off-peak pricing by UTC window makes budgeting and benchmarking fiddly
- Concurrency caps (2,500 for Flash, 500 for V4-Pro) and periodic capacity squeezes hit production workloads
- The chat product trails ChatGPT, Gemini and Claude on connectors, projects and team features
Best for
- Developers and agent builders who want frontier-class output at a fraction of OpenAI or Anthropic token prices
- Teams that want open weights they can host themselves and swap in behind the same API shape
- Individuals who want a free reasoning assistant with a 1M-token context and no subscription
Not for
- Organisations whose policies forbid sending prompts to servers in China, including several governments that have banned the app on official devices
- Teams needing enterprise admin, SSO, audit logs or a signed data-processing agreement for the chat product
Watch out: The chat app stores conversations on servers in China and the API answers to PRC law; the headline API prices are also off-peak rates, so a workload that runs during European mornings pays double.
Alternatives to DeepSeek
DeepSeek head-to-head
Side-by-side fact sheets against the competitors our research or their makers name — pricing, platforms, API and both verdicts on one page.