Provider pricing hub
DeepSeek API Pricing (July 2026)
The DeepSeek API costs $1.74/$3.48 per million input/output tokens for DeepSeek V4 Pro and $0.14/$0.28 for V4 Flash, with cache-hit input as low as $0.028 — one to two orders of magnitude cheaper than Western flagships. V3.2 remains available at $0.28/$0.42. Both V4 tiers ship a 1M-token context window.
DeepSeek is the price floor of the frontier-adjacent API market. The V4 generation kept the strategy that made V3 famous: one flagship (V4 Pro, $1.74/$3.48) priced below Western mid-tiers, one fast model (V4 Flash, $0.14/$0.28) priced below almost everything, and automatic context caching that cuts repeated input to cents. The table below lists every DeepSeek model in BenchLM's registry, synced from DeepSeek's official pricing page.
Cheap only matters if the quality clears your bar: see the Chinese model rankings (where the V4 Pro reasoning rows lead), the price-vs-performance view, and the Chinese LLM deep dive.
Every DeepSeek API price per 1M tokens
| Model | Input $/M | Cached input $/M | Output $/M | Context |
|---|---|---|---|---|
| DeepSeek V4 Pro | $1.74 | $0.174 | $3.48 | 1M |
| DeepSeek R1 | $0.55 | $0.055 | $2.19 | 128K |
| DeepSeek V3.2 (Thinking) | $0.55 | $0.055 | $2.19 | 128K |
| DeepSeek V3.2 | $0.28 | $0.028 | $0.42 | 128K |
| DeepSeek V3 | $0.27 | $0.027 | $1.1 | 128K |
| DeepSeek V4 Flash | $0.14 | $0.014 | $0.28 | 1M |
| DeepSeek LLM 2.0 | $0 | $0 | $0 | 128K |
| DeepSeek R1 Distill Qwen 32B | $0 | $0 | $0 | 128K |
| DeepSeek V3.1 | $0 | $0 | $0 | 128K |
| DeepSeekMath V2 | $0 | $0 | $0 | 128K |
Cache-hit input rates are DeepSeek's official published prices (cache misses bill at the standard input rate). DeepSeek does not currently offer a Batch API discount tier; reasoning-effort variants (High/Max) inherit their family's token pricing and are collapsed into one row.
Estimate your monthly DeepSeek API bill
| Model | Est. monthly cost |
|---|---|
| DeepSeek V4 Pro | $24.36 |
| DeepSeek R1 | $9.88 |
| DeepSeek V3.2 (Thinking) | $9.88 |
| DeepSeek V3.2 | $3.64 |
| DeepSeek V3 | $4.9 |
| DeepSeek V4 Flash | $1.96 |
| DeepSeek LLM 2.0 | $0 |
| DeepSeek R1 Distill Qwen 32B | $0 |
Estimates use DeepSeek's standard per-token rates from the table above. For task-level presets and cross-provider comparison, use the full AI cost calculator.
Cache-hit pricing: DeepSeek’s biggest discount is automatic
DeepSeek prices input tokens in two lanes: cache miss (the standard rate) and cache hit, billed when your prompt prefix was recently seen. On V4 Flash and V3.2 a cache hit costs $0.028 per million tokens— effectively free — and on V4 Pro it's $0.145. Caching is automatic on the hosted API: no cache-control headers, no write surcharge, no minimum TTL to manage.
For agents, RAG pipelines, and chat applications that resend the same system prompt or document context, the cached share often reaches 60–80% of input volume. At that mix, V4 Pro's effective input rate lands near $0.50/M — model it with the estimator above. There is no Batch API; DeepSeek's position is that the list price already is the discount.
V4 Pro vs V4 Flash: which DeepSeek tier to use
V4 Pro($1.74/$3.48) is the flagship: its Max reasoning-effort row scores 75 on BenchLM's provisional leaderboard and leads our Chinese model rankings, with a 1M-token context window. Use it for coding, reasoning, and agentic work where you'd otherwise reach for a Western mid-tier at 2–3x the price. V4 Flash ($0.14/$0.28) is the volume tier — classification, extraction, summarization, and high-throughput pipelines where per-request cost dominates.
The older V3.2 ($0.28/$0.42 non-thinking, $0.55/$2.19 thinking) and R1 remain on the price sheet, and the V4 family is open weight — you can self-host instead of paying per token.
DeepSeek vs OpenAI and Claude pricing
The gap is not subtle. V4 Pro at $1.74/$3.48 costs roughly a third of GPT-5.6 Terra ($2.50/$15) on output and ~9x less than GPT-5.6 Sol or Claude Opus 4.8 ($30 and $25 output respectively). V4 Flash at $0.14/$0.28 undercuts even OpenAI's nano tier on output. The honest caveat: the Western flagships still score higher on BenchLM's overall leaderboard, so the decision is quality-per-dollar, not price alone.
Compare the full tables: OpenAI API pricing and Claude API pricing, or the head-to-head Opus 4.8 vs V4 Pro comparison on the compare hub.
DeepSeek API pricing FAQ
How much does the DeepSeek API cost?
DeepSeek V4 Pro costs $1.74 input / $3.48 output per million tokens and V4 Flash costs $0.14/$0.28, with cache-hit input at $0.145 and $0.028 respectively. The older V3.2 endpoint remains available at $0.28/$0.42. There are no separate batch tiers — the standard rates are already the discount.
Is the DeepSeek API free?
No — the hosted API is pay-per-token, though at prices low enough that light usage costs cents per day. The models themselves are open weight, so the genuinely free path is self-hosting a distilled or quantized variant on your own hardware; see the open-source model rankings.
What is DeepSeek cache-hit pricing?
When your request reuses a recently seen prompt prefix, DeepSeek bills that portion at the cache-hit rate instead of the standard input rate — $0.028/M on V4 Flash and V3.2, $0.145/M on V4 Pro. It is automatic on the hosted API and typically covers most input tokens in agent and RAG workloads.
How does DeepSeek API pricing compare to OpenAI?
DeepSeek V4 Pro's output tokens cost about 9x less than GPT-5.6 Sol's ($3.48 vs $30 per million), and V4 Flash undercuts GPT-5.6 Luna by ~21x on output ($0.28 vs $6). OpenAI's flagships still lead on BenchLM's overall scores, so the right comparison is quality-per-dollar for your specific workload — see the OpenAI pricing hub.