DeepSeek V4 Flash pricing: the cheapest named LLM, costed
DeepSeek V4 Flash is the cheapest named model in the live OpenRouter catalog: $0.012 per 1M input tokens, $1.25 per 1M output, with a 1,310,720-token context window. Here’s the full breakdown, including what heavy daily use actually costs.
Every DeepSeek model, priced
All DeepSeek rows on OpenRouter, cheapest input first.
| Model | Context window | Input /1M | Output /1M |
|---|---|---|---|
| DeepSeek V4 Flash Latest | 1,310,720 | $0.012 | $1.25 |
| DeepSeek V4 Flash 0731 | 1,310,720 | $0.018 | $0.32 |
| DeepSeek Flash Latest | 1,048,576 | $0.02 | $0.6 |
| DeepSeek V4 Flash 0423 | 1,048,576 | $0.076 | $0.153 |
| DeepSeek V4.1 Flash (batch) | 1,048,576 | $0.112 | $0.336 |
| DeepSeek Pro Latest | 1,048,576 | $0.15 | $3.50 |
| DeepSeek V4 Flash Vision Exp | 1,048,576 | $0.216 | $0.647 |
| DeepSeek V3.1 | 163,840 | $0.25 | $0.95 |
| DeepSeek V3 | 163,840 | $0.257 | $1.03 |
| DeepSeek V3.2 Exp | 163,840 | $0.27 | $0.41 |
| DeepSeek V3.2 | 163,840 | $0.28 | $0.42 |
| DeepSeek V3 0324 | 163,840 | $0.29 | $1.14 |
| DeepSeek V4.1 Flash | 1,048,576 | $0.3 | $1.20 |
| DeepSeek V3.1 Terminus | 163,840 | $0.3 | $1.00 |
| DeepSeek V4 Pro 0813 | 1,048,576 | $0.475 | $4.20 |
| DeepSeek R1 0528 | 163,840 | $0.5 | $2.15 |
| DeepSeek R1 | 64,000 | $0.7 | $2.50 |
| DeepSeek V4 Pro 0423 | 1,048,576 | $0.936 | $1.87 |
What heavy use costs per month
Scenario: 1M input tokens and 250K output tokens every day for 30 days. This is roughly a production app doing constant summarization or RAG over long documents.
| Model | Monthly cost |
|---|---|
| DeepSeek V4 Flash Latest | $9.73 |
| DeepSeek V4 Flash 0731 | $2.94 |
| Qwen3.7 Flash | $1.88 |
| GPT-4o-mini | $9.00 |
DeepSeek V4 Flash runs that workload for less than 2% of what GPT-4o Mini would charge. Same traffic, $0.92x smaller bill.
DeepSeek Flash vs the other budget leaders
| Model | Input /1M | Output /1M | Context window |
|---|---|---|---|
| DeepSeek V4 Flash Latest | $0.012 | $1.25 | 1,310,720 |
| GPT-4o-mini | $0.15 | $0.6 | 128,000 |
| Qwen3.7 Flash | $0.03 | $0.13 | 1,000,000 |
Output pricing is where DeepSeek V4 Flash is unusual: $1.25 is actually higher than the 0731 sibling’s $0.32. If your workload is generation-heavy, check both rows — the older sibling can be cheaper overall.
Why is it so cheap?
DeepSeek prices aggressively across the board and trains efficient flash-tier models designed for high-volume serving. Whatever the internal economics, the published rates are the lowest in the catalog — and they’re live rates you can verify on any DeepSeek row in the OpenRouter pricing calculator.
FAQ
What is the cheapest LLM available right now?
In the live OpenRouter catalog, DeepSeek V4 Flash at $0.012 per 1M input tokens is the cheapest named model. See the full ranking in our cheapest 1M-context LLMs page.
How much does 1M output tokens cost on DeepSeek V4 Flash?
$1.25. Input is the bargain here — if your workload generates a lot, compare the $0.32 output rate on the 0731 variant.
Is the low price reliable for production?
The price is live in the catalog today. Flash-tier models trade peak reasoning quality for cost and speed, so benchmark on your own tasks before routing production traffic.
More pricing breakdowns
Browse the answers index, or compare with Qwen pricing 2026 and GPT-5.5 Pro pricing.