> DeepSeek
DeepSeek's official API serves the current V4.1-Flash and V4-Pro models with 1M-token context and 384K max output, from $0.15/$0.60 per 1M in/out off-peak (Flash) — peak hours bill at 2x, cached input drops to $0.003/M. One key exposes both an OpenAI-compatible and an Anthropic-compatible endpoint, and sign-up grants a free token balance that is deducted before any top-up.
// pros
- Among the cheapest frontier-class open-model pricing available: Flash input (cache miss) is $0.15/M off-peak, and cache hits fall to $0.003/M.
- 1M context and 384K max output on both current models — long-document work fits without chunking.
- Two API surfaces from one key: OpenAI-compatible and Anthropic-compatible endpoints, plus tool calls, JSON mode and a Responses API.
- Off-peak hours (everything except Mon–Fri 01:00–04:00 and 06:00–10:00 UTC) bill at exactly half the peak rate — batch jobs can halve their bill by when they run.
- The free 'granted balance' from sign-up is deducted before any topped-up balance — no separate free endpoint to manage.
// cons
- The same token costs 2x at peak vs off-peak — production traffic without off-peak scheduling bills at the peak rate.
- Concurrency is capped per model (2,500 Flash / 500 V4 Pro) — high-fanout workloads can hit the ceiling.
- Vision is Flash-only; V4-Pro takes no image input.
- Model churn: legacy names (e.g. deepseek-v4-flash) are retired and silently served by the Flash model at Flash prices — pin model names deliberately.
The cheapest credible scale option in this directory: Flash for high-volume, off-peak-capable work; V4-Pro when quality outweighs the price delta. Turn caching on — at $0.003/M for Flash cache hits, repeated prefixes are effectively free.
written by Mike Fleming, Uprouter editorial · verified against official sources September 16, 2026
How we estimated this: the free allowance is converted to a dollar value using the provider’s own pay-as-you-go rates for the models a typical developer would reach for. When the allowance is metered in requests, tokens or minutes rather than dollars, we assume median usage of the cheapest capable model. Because assumptions are involved, the figure carries the est. flag — treat it as an order of magnitude, not a quote.
note: Free 'granted balance' of ~5M tokens at sign-up (deducted before top-ups per official billing rules); $0.70 is an estimate at V4-Flash off-peak rates.
| Plan | Type | Monthly | Input /1M | Output /1M | Note |
|---|---|---|---|---|---|
| Free tier | free | — | — | — | Free tier (source-reported) |
| Pay-as-you-go | payg | — | Free | $0.0000 | Cheapest tracked model rate (OpenRouter-normalized). |
| DeepSeek-V4-Flash (Peak) | payg | — | $0.140 | $0.280 | Peak hours apply (UTC 01:00-04:00, 06:00-10:00). Off-peak is 50% cheaper. |
| DeepSeek-V4-Pro (Peak) | payg | — | $0.440 | $1.32 | Reasoning/Thinking model. |
| Free (web chat) | free | — | — | — | DeepSeek web chat is free; a separate pay-as-you-go API exists. |
Pricing normalized from public sources — always verify with the provider.
// model prices
| Model | Input /1M | Output /1M | Context |
|---|---|---|---|
| DeepSeek Chat | Free | $0.0000 | 164k |
| DeepSeek-R1 | $0.0000 | $0.0000 | 64k |
| DeepSeek V4 Flash Vision Exp | $0.220 | $0.660 | 1.0M |
| DeepSeek V3.1 | $0.550 | $1.65 | 161k |
| DeepSeek V4 Pro 0813 | $1.12 | $3.36 | 1.0M |
| deepseek-v4-pro | — | — | 1.0M |
| DeepSeek V4 Flash | — | — | 1.0M |
| DeepSeek V3.2 | — | — | 164k |
| DeepSeek: DeepSeek Coder V2 Instruct | — | — | 128k |
$ Does DeepSeek have a free tier?
Yes — DeepSeek offers a free tier that we value at roughly $0.70 in free credits (estimate — derived from rate-limited usage at pay-as-you-go prices). Always confirm current limits on DeepSeek's official pricing page — free tiers change without notice.
$ How much does DeepSeek cost per 1M tokens?
The cheapest model we track at DeepSeek is Free per 1M input tokens and $0.0000 per 1M output tokens. This is normalized from DeepSeek's published pricing — verify with the provider before purchasing, since prices change frequently.
$ Is DeepSeek safe to use?
Uprouter rates DeepSeek as medium risk. Low risk: a first-party API from a named model lab with a single authoritative pricing page that states peak/off-peak rules, model deprecations and billing precedence explicitly. The main exposures are schedule-driven cost (identical tokens billed at 2x during peak hours) and silent re-routing when legacy model names are retired. Risk ratings are editorial, evidence-based, and never influenced by affiliate relationships.
Raw community sentiment, shown unblended. Votes never feed into objective data fields like price, risk, or status — and never into sort order.
No published reviews yet — be the first.
// more_informationprice history · live status · risk & confidence · code snippets
price_history
No recorded changes yet.
live_status
current: unknownNo probes recorded yet.
Uprouter runs its own probes; best-effort, not a guarantee. Probe results depend on our network location and cadence — for production, always use the provider’s own status page.
risk_&_confidence
Low risk: a first-party API from a named model lab with a single authoritative pricing page that states peak/off-peak rules, model deprecations and billing precedence explicitly. The main exposures are schedule-driven cost (identical tokens billed at 2x during peak hours) and silent re-routing when legacy model names are retired.
Confidence reflects how much verifiable evidence (official docs, probe history, community reports) backs this entry. Risk is our editorial assessment of operator reliability and terms — not financial advice.
code_snippets
# DeepSeek's own API — get a key from their dashboard
curl https://chat.deepseek.com/chat/completions \
-H "Authorization: Bearer THEIR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek/deepseek-chat-v3.1",
"messages": [{ "role": "user", "content": "Hello" }]
}'Direct to the provider — Uprouter receives nothing; verify their current pricing before purchasing.
Ranked by neutral similarity signals only — live status, free-tier shape and free-credit proximity. Commercial relationships never influence this list.