Cheap DeepSeek API: Flash vs Pro Prices and API Key Access
DeepSeek Flash from $0.18/$0.72 and V4 Pro from $0.9504/$2.8512 per 1M tokens at the peak-time check. Compare discounts, cache and API key access.
Published
DeepSeek Flash starts at $0.18 input / $0.72 output; V4 Pro starts at $0.9504 / $2.8512. At this peak-time snapshot the lowest routes are 40% and 28% below the corresponding official peak prices. Buy by exact model and billing window.
Checked 2026-11-04 03:32 UTC. Prices are USD per 1M input / output tokens; the table shows the lowest and highest input-priced public route as input / output pairs. Counts exclude client-only routes. The live marketplace card controls current price and availability.
| Model ID | Official reference price | APIMaster current price | Active routes | Discount | Model card |
|---|---|---|---|---|---|
deepseek-flash |
$0.3 / $1.2 | $0.18 / $0.72 – $0.3 / $1.2 | 3 | Up to 40% off | Compare routes |
deepseek-v4-pro |
$1.32 / $3.96 | $0.9504 / $2.8512 – $6.400944 / $19.202832 | 5 | Up to 28% off | Compare routes |
Snapshot taken during the November 4 peak window. Official Flash peak/off-peak is $0.30/$1.20 versus $0.15/$0.60; Pro is $1.32/$3.96 versus $0.66/$1.98. These are not all-day fixed quotes. Some listed Pro routes exceed official Standard prices.
GPT Trial(GPT-6 Astra, GPT-6 Sol, GPT-5.6, GPT-5.5, gpt-5.6-luna, gpt-5.6-sol, gpt-5.6-terra, gpt-6.1-sol, gpt-6-astra, gpt-6-sol, and gpt-6-luna available)
Sign up and claim a $20 GPT trial
DeepSeek API pricing: official vs OpenRouter vs APIMaster
| Provider / model | Input / output | Billing condition | APIMaster saving |
|---|---|---|---|
| APIMaster · deepseek-flash | $0.18 / $0.72 | Standard, uncached | — |
| DeepSeek Flash · Peak | $0.30 / $1.20 | Peak, uncached | At least 40% lower token rates |
| DeepSeek Flash · Off-peak | $0.15 / $0.60 | Off-peak; different time window | — |
| OpenRouter · deepseek/deepseek-v4.1-flash-1017 | $0.30 / $1.20 | Peak, uncached | At least 40% lower token rates |
| OpenRouter · deepseek/deepseek-v4.1-flash | $0.045 / $0.50 | Different alias/provider selection | — |
Peak hours are Monday–Friday 01:00–04:00 and 06:00–10:00 UTC, excluding Chinese public holidays; all other hours are off-peak. Official Flash cache reads are $0.006 peak / $0.003 off-peak; Pro $0.044 / $0.022. The OpenRouter plain Flash alias has lower listed rates than this APIMaster snapshot; its 1017 alias and Batch are separate listings. We do not claim APIMaster wins every rate.
What does a real token budget cost?
Budget: 100,000 uncached input tokens + 20,000 billed output tokens per task. Formula: input_tokens × input_rate / 1,000,000 + output_tokens × output_rate / 1,000,000. This is arithmetic, not a measured model benchmark.
| Model ID | Official reference / task | APIMaster from / task | APIMaster / 1,000 tasks |
|---|---|---|---|
deepseek-flash |
$0.054 | $0.0324 | $32.4 |
deepseek-v4-pro |
$0.2112 | $0.152064 | $152.06 |
The 1,000-task column scales the same budget; it is not a bulk discount. Figures are rounded and exclude cache, storage, search/tools, taxes and retries. Include billed reasoning output and every agent call. Different models can consume different token counts; compare cost per accepted result.
Which DeepSeek model or route should you buy?
Start with Flash for the smaller token bill and evaluate Pro only if its behavior improves your accepted results. At this peak snapshot, 1,000 Flash tasks at the table's budget cost $32.40 versus $54 at official peak pricing. Recheck the card when the time window changes.
An old screenshot with half these prices can describe off-peak hours rather than an expired promotion. Compare identical timestamps, cache status and model aliases. Search/Fetch, reasoning output, retries and client-side agent loops can change the bill; topping up APIMaster does not recharge a DeepSeek account.
Choose the route inside the same model card and confirm supported tools, streaming and limits. Fingerprint results are route-specific identity evidence; availability records are point-in-time checks, not an uptime guarantee. Missing test records do not count as passes.
How to get a discounted DeepSeek API key
Compare the live model card → register → fund your wallet → create an API key. Use your APIMaster key, https://apimaster.ai/v1, and the exact model ID deepseek-flash:
curl "https://apimaster.ai/v1/chat/completions" \
-H "Authorization: Bearer $APIMASTER_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"deepseek-flash","max_tokens":512,"messages":[{"role":"user","content":"Return a JSON object with a short summary of: API costs depend on token usage."}]}'
This is a minimal integration example, not a new paid inference test. Check endpoint compatibility, returned usage and your first small bill before increasing traffic. An APIMaster key does not fund an account at the original vendor.
FAQ
Is this the cheapest API on the internet?
No universal lowest-price claim is made. This guide compares the checked public routes and named billing modes; another provider or specialized route can be cheaper.
Does an app subscription include these API credits?
Treat app subscriptions, coding plans and pay-as-you-go API wallets as separate products unless their terms explicitly include the same usage.
Does the headline discount also cover cache hits?
No. Compare cached reads, cache writes, storage, tools and timing separately; ordinary token discounts do not prove the same saving on every billing item.
Does the GPT trial promotion provide this model for free?
The GPT trial offer is separate. It does not promise free credits for the models in this guide; check the current trial catalog and eligibility.
Sources and related guides
Official pricing · OpenRouter model catalog
DeepSeek V4.1 Flash · DeepSeek V4 Pro · GLM FlashX vs DeepSeek Flash
Other cheap API guides: Claude Sonnet 5.5 · Gemini Flash · GLM · Qwen
Get started: Compare the live model card → register → fund your wallet → create an API key. Use your APIMaster key, https://apimaster.ai/v1, and the exact model ID deepseek-flash