DeepSeek V4.1 Flash API pricing
DeepSeek · Budget price tier · MIT · released 2026-09. Prices in USD per 1 million tokens. Data snapshot: 2026-10-26.
Outlier listing. The price we quote is one host's rate with input near free and output at or above the going rate, not what most hosts charge, so a cost built on it can be far from what you would pay. The typical host charges about $0.19 per 1M input tokens for this model.
A support ticket workload of 100K requests a month costs $53.00 on DeepSeek V4.1 Flash; the next cheaper model in the Budget price tier for this workload, Qwen3 VL 30B A3B Instruct, does it for $52.50. The price we quote for DeepSeek V4.1 Flash has changed 25 times since 2026-09-11, most recently on 2026-10-24 when the quoted host changed (-24.1%); it is now the 109th cheapest of 164 models in the Budget price tier. Buying from Decart rather than at the price we quote lowers the token price by about 78%, for a quantized fp4 build. It is rated a stable release but does not meet all four criteria of the FinOps Friendly badge.
- Input: $0.02 per 1M tokens
- Output: $1.00 per 1M tokens
- Batch API (in/out): Not published
- Prompt-cache read: $0.015 per 1M tokens
- Context window: 1M tokens
- Arena ELO: 1475 as of 2026-10-08
Market position
On blended price ($0.706 per 1M tokens at 30% input, 70% output) it is cheaper than 34% of models in the Budget price tier: 109th cheapest of 164. Across every price tier it is cheaper than 65% of priced text models in the catalogue: 109th cheapest of 310. Its Arena ELO of 1475 ranks 25th of the 156 models that carry one, tied with 5 others, above 81% of them. With an Arena score of 1475, DeepSeek V4.1 Flash clears OptimNow's Strong and Standard working quality thresholds: Strong (Arena 1465+), used for support ticket, knowledge Q&A and marketing content; Standard (Arena 1415+), used for meeting summary, invoice processing and call summary. Each threshold is a distance below the best ranked model, which scores 1525 on Arena scores as of 8 Oct 2026.
Licence, size and capabilities
DeepSeek V4.1 Flash is open source: its weights are published under MIT, an OSI-approved licence with no field-of-use limit, so it can be self-hosted as well as bought by the token. Its parameter count is not published. Listed capabilities: Text, Vision, Reasoning, Agents. Filling its 1M context window once costs $0.02 in input tokens at list price.
Pricing by host
A host is a company that runs the model and sells it by the token: the model maker, a cloud platform like Amazon Bedrock, Azure or Vertex, or an independent host. One model, several hosts, and often several prices. This model is sold by 33 different hosts. The price we quote elsewhere on this page is marked. We quote one host's rate so that models stay comparable; it is not a recommendation to buy there.
Cheapest host: Decart at $0.09 input / $0.18 output per 1M tokens, 78% below the price we quote on the blended price (30% input, 70% output). It serves a quantized fp4 build, which is not the same product as the full-precision model. A support ticket workload at 100K requests a month costs $22.50 there against $53.00 at the price we quote, a difference of $30.50 a month. The dearest, Venice, charges 6.1x the cheapest.
| Host | Input $/1M | Output $/1M | Precision |
|---|---|---|---|
| Decart | $0.09 | $0.18 | fp4 |
| Sail Research | $0.09 | $0.25 | fp4 |
| DeepInfra | $0.14 | $0.42 | fp8 |
| Relace | $0.07 | $0.50 | not declared |
| AtlasCloud | $0.135 | $0.54 | fp8 |
| Baidu | $0.1398 | $0.5592 | fp8 |
| StreamLake | $0.147 | $0.588 | fp8 |
| InferenceNet | $0.103 | $0.645 | fp8 |
| CoreWeave | $0.15 | $0.65 | fp8 |
| DigitalOcean | $0.165 | $0.66 | not declared |
| Alibaba | $0.165 | $0.66 | not declared |
| Wafer | $0.19 | $0.65 | mxfp4 |
| ModelRun | $0.17 | $0.67 | fp4 |
| GMICloud | $0.18 | $0.72 | fp8 |
| Morph | $0.098 | $0.80 | fp8 |
| Novita | $0.195 | $0.78 | fp8 |
| Reka | $0.105 | $0.84 | not declared |
| Fireworks | $0.21 | $0.84 | not declared |
| Phala | $0.21 | $0.84 | not declared |
| OpenInference (price we quote) | $0.02 | $1.00 | fp4 |
| Mancer 2 | $0.20 | $1.00 | fp8 |
| Makora | $0.26 | $1.17 | fp8 |
| Crusoe | $0.29 | $1.20 | fp8 |
| $0.30 | $1.20 | not declared | |
| Nebius | $0.30 | $1.20 | fp8 |
| PrimeIntellect | $0.30 | $1.20 | not declared |
| Cohere | $0.30 | $1.20 | not declared |
| BaseTen | $0.30 | $1.20 | not declared |
| Together | $0.30 | $1.20 | not declared |
| SiliconFlow | $0.30 | $1.20 | fp8 |
| Modal | $0.30 | $1.20 | not declared |
| Parasail | $0.30 | $1.20 | fp8 |
| Venice | $0.30 | $1.20 | fp8 |
List prices as published on 2026-10-26, via OpenRouter. Negotiated rates, committed-use discounts and provisioned capacity are not shown. These hosts do not all serve the model at the same numeric precision, so the cheapest is not a like-for-like substitute for the dearest. These listings are promotional rates, below the host's standard rate: DeepInfra's, 30% off; Baidu's, 53% off; StreamLake's, 51% off; GMICloud's, 40% off; Novita's, 35% off; Reka's, 30% off; and Phala's, 30% off.
Monthly cost at 100K requests, by business use case
From the after-discounts figure (prompt caching on the cached input share, batch API for async workloads, only where DeepSeek publishes those rates), which assumes you implement both, to the list price. One figure means no discount is published.
| Use case | Tokens (in/out) | Monthly (after discounts – list) |
|---|---|---|
| Support Ticket | 1,500 / 500 | $53 |
| Knowledge Q&A | 2,000 / 800 | $83 – $84 |
| Meeting Summary | 10,000 / 1,200 | $140 |
| Marketing Content | 2,500 / 1,800 | $185 |
| Coding Task | 3,000 / 2,000 | $205 – $206 |
| Invoice Processing | 1,500 / 600 | $63 |
| Call Summary | 2,000 / 700 | $74 |
| Agent Workflow | 6,000 / 3,000 | $310 – $312 |
Price history
DeepSeek V4.1 Flash entered our daily price archive on 2026-09-11; tracking began on 2026-08-15. It listed at $0.15 input / $0.60 output per 1M tokens that day. The price we quote has changed 25 times since then, 16 of them a change of quoted host rather than a repricing. The latest change, on 2026-10-24, was -24.1% on the blended price (30% input, 70% output), to $0.02 input / $1.00 output (quoted host changed, not a repricing). Net of all changes, the blended price is +51.8% against 2026-09-11.
| Date | Old price (in / out) | New price (in / out) | Blended change |
|---|---|---|---|
| 2026-10-24 | $0.30 / $1.20 | $0.02 / $1.00 | -24.1% (quoted host changed, not a repricing) |
| 2026-10-22 | $0.049 / $0.375 | $0.30 / $1.20 | +235.5% (quoted host changed, not a repricing) |
| 2026-10-21 | $0.056 / $0.12 | $0.049 / $0.375 | +175.0% |
| 2026-10-20 | $0.0376 / $0.08 | $0.056 / $0.12 | +49.8% |
| 2026-10-19 | $0.04 / $0.144 | $0.0376 / $0.08 | -40.4% |
20 earlier changes are not listed.
Alternatives to DeepSeek V4.1 Flash
2 of the 4 comparable models below cost less than DeepSeek V4.1 Flash: GLM 5.3 Flash undercuts it by 44%, at the same Arena score.
- Next cheaper model in the Budget price tier: Qwen3.6 35B A3B (Alibaba) at $0.695 blended, 2% less.
- Next more expensive: Codestral 2508 (Mistral) at $0.72 blended, 2% more.
The nearest models in the Budget price tier on Arena ELO, each with its blended price and score difference against DeepSeek V4.1 Flash.
- GLM 5.3 Flash (Zhipu): $0.15 in / $0.50 out per 1M, Arena ELO 1475 (44% cheaper, same Arena ELO)
- GLM 5.3 (Zhipu): $0.0699 in / $1.40 out per 1M, Arena ELO 1478 (42% more expensive, 3 ELO points higher)
- MiMo-V2.6-Pro (Xiaomi): $0.435 in / $0.87 out per 1M, Arena ELO 1480 (5% more expensive, 5 ELO points higher)
- MiMo-V2.5-Pro (Xiaomi): $0.3045 in / $0.609 out per 1M, Arena ELO 1468 (27% cheaper, 7 ELO points lower)
- Cheapest within 15 Arena points: GLM 5.3 Flash (Zhipu), 44% cheaper, same Arena ELO, out of 32 models in that band.
- Highest Arena ELO within 20% of this price: MiMo-V2.6-Pro (Xiaomi), 5% more expensive, 5 ELO points higher, out of 13 models in that price band.
Other models from DeepSeek
The 8 nearest to DeepSeek V4.1 Flash in price, of 13 other DeepSeek models in the catalogue. Prices per 1M tokens, input then output.
- DeepSeek V3.2 Exp: $0.27 / $0.41
- DeepSeek V3.2: $0.28 / $0.42
- DeepSeek V4.1 Flash 1017: $0.15 / $0.60
- DeepSeek V4 Flash Vision Exp: $0.2156 / $0.6468
- DeepSeek V3.1: $0.25 / $0.95
- DeepSeek V3.1 Terminus: $0.27 / $1.00
- DeepSeek V3: $0.2574 / $1.0287
- DeepSeek V3 0324: $0.29 / $1.14
All 14 DeepSeek models and prices
Common questions about DeepSeek V4.1 Flash pricing
- How much does DeepSeek V4.1 Flash cost?
- DeepSeek V4.1 Flash costs $0.02 per 1M input tokens and $1.00 per 1M output tokens at OpenInference, the host whose rate we quote. OpenInference serves a quantized fp4 build, which is not the same product as the full-precision model. Prices as listed on 2026-10-26.
- Which host is cheapest for DeepSeek V4.1 Flash?
- Cheapest host: Decart at $0.09 input / $0.18 output per 1M tokens, 78% below the price we quote on the blended price (30% input, 70% output). It serves a quantized fp4 build, which is not the same product as the full-precision model. A support ticket workload at 100K requests a month costs $22.50 there against $53.00 at the price we quote, a difference of $30.50 a month. The dearest, Venice, charges 6.1x the cheapest. Host rates as listed on 2026-10-26.
- Does DeepSeek V4.1 Flash offer batch or cached pricing?
- Yes, for prompt caching only. Cache reads cost $0.015 per 1M tokens, as listed on 2026-10-26. No batch pricing is published.
- What would DeepSeek V4.1 Flash cost per month?
- With the Support Ticket profile (1,500 input and 500 output tokens per request), 100,000 requests a month cost about $53 on DeepSeek V4.1 Flash at the list price quoted on 2026-10-26.
- Can DeepSeek V4.1 Flash be self-hosted?
- Yes. DeepSeek V4.1 Flash is open source: its weights are published under MIT, an OSI-approved licence with no field-of-use limit, so it can be self-hosted as well as bought by the token.
- Has the price of DeepSeek V4.1 Flash changed?
- DeepSeek V4.1 Flash entered our daily price archive on 2026-09-11; tracking began on 2026-08-15. It listed at $0.15 input / $0.60 output per 1M tokens that day. The price we quote has changed 25 times since then, 16 of them a change of quoted host rather than a repricing. The latest change, on 2026-10-24, was -24.1% on the blended price (30% input, 70% output), to $0.02 input / $1.00 output (quoted host changed, not a repricing). Net of all changes, the blended price is +51.8% against 2026-09-11.
Its weights are published, which changes what you may run yourself: the cost gap between open weights and proprietary APIs.
These are list prices. Your real cost also depends on caching and batching, negotiated discounts, marketplace agreements, and the contracts you already have. OptimNow audits exactly that.
Build the business case · Get a FinOps review
LLM pricing · Models by maker · Cloud compute pricing · Compare models · Price barometer · FinOps guides · API and MCP server