DeepSeek V4.1 Flash 1017
DeepSeekReleased Oct 17, 2026Open weights · MITdeepseek-v4.1-flash-1017
- Reasoning
- yes, with effort levels
- Context
- 1M tokens
- Max output
- 384K tokens
- Input
- text, image
- API features
- tools, structured outputs, JSON mode, seed, logprobs
- DeepSeek's API
- yes, as deepseek-flash
Read Oct 23, 06:00 UTC from the hosts' published offers and OpenRouter, spec sheet from DeepSeek
Speed by provider
Output speed of DeepSeek V4.1 Flash 1017 at each host, fastest first (the 10 fastest of 20)
- Together AI268 tok/s0.31 s to first token
- BaseTen250 tok/s0.34 s to first token
- DigitalOcean197 tok/s0.73 s to first token
- Morph162 tok/s0.55 s to first token
- Nebius156 tok/s1.22 s to first token
- Relace124 tok/s0.97 s to first token
- Fireworks AI121 tok/s1.04 s to first token
- SiliconFlow119 tok/s2.53 s to first token
- Crusoe117 tok/s0.61 s to first token
- AtlasCloud113 tok/s1.20 s to first token
Speed: OpenRouter, median of the last 30 min, read Oct 23, 06:00 UTC · a snapshot of real traffic before the daily read, not a daily average · bars from zero.
- Parameters
- 552B
- 16B active per token (MoE)
- Memory needed
- 331 GB
- in INT4 · 662 GB in INT8
- Smallest cloud setup
- 8× H100
- 640 GB · $17.20/h, about $12,556 a month
- Worth it above
- 32B tokens
- a month, against the cheapest API at $0.39 per M
| Quantization | Memory | Smallest cloud GPU | Quality |
|---|---|---|---|
| INT4 (Q4) | 331 GB | 8× H100 · $17.20/h | smallest footprint, minor quality loss |
| INT8 (Q8) | 662 GB | multi-node, over 640 GB | near-lossless |
| FP16 (full) | 1325 GB | multi-node, over 640 GB | reference quality |
At home: 331 GB in INT4 is beyond any desktop or Mac in our machines table, so it means a multi-GPU server.
Memory ≈ parameters × bytes per parameter × 1.2 · cloud price: on-demand rate, running 24/7 · builds: GGUF, AWQ, MLX, GPTQ, Ollama · break-even calculator
| Provider | Input | Output | Cache ↑ | tok/s | First token | Uptime 24h | Checked |
|---|---|---|---|---|---|---|---|
| DeepInfrafp8 · 1M ctx · 128K out−51% | $0.098 | $0.29 | $0.0029 | 67 | 1.03 s | 100% | Oct 23 |
| Cloudflare1M ctx · 375K out | $0.10 | $0.60 | $0.003 | 85 | 1.79 s | 99.9% | Oct 23 |
| DeepSeek1M ctx · 384K outoff-peak−50% | $0.15 | $0.60 | $0.003 | — | — | — | Oct 22 |
| Rekaprecision not read: host terms · 1M ctx · 384K out | $0.15 | $0.60 | $0.003 | 104 | 0.79 s | 97.2% | Oct 23 |
| GMICloudfp8 · 1.05M ctx · 375K out−40% | $0.18 | $0.72 | $0.0036 | 98 | 1.97 s | 99.9% | Oct 23 |
| AtlasCloudfp8 · 1M ctx · 384K out | $0.20 | $0.78 | $0.0039 | 113 | 1.20 s | 99.0% | Oct 23 |
| Modal1M ctx | $0.25 | $1.00 | $0.005 | 89 | 0.83 s | 98.3% | Oct 23 |
| Baidufp8 · 1M ctx · 384K out | $0.28 | $1.13 | $0.006 | 97 | 1.22 s | 97.9% | Oct 23 |
| Phalaprecision not read: host terms · 1M ctx · 384K out | $0.30 | $1.20 | $0.006 | 111 | 1.13 s | 98.2% | Oct 23 |
| SiliconFlowfp8 · 1M ctx · 384K out | $0.30 | $1.20 | $0.006 | 119 | 2.53 s | 99.2% | Oct 23 |
| Together AIfp8 · 1M ctx | $0.30 | $1.20 | $0.006 | 268 | 0.31 s | 100% | Oct 23 |
| Fireworks AI1M ctx · 375K out | $0.30 | $1.20 | $0.006 | 121 | 1.04 s | 98.2% | Oct 23 |
| Parasailfp8 · 1M ctx · 375K out | $0.30 | $1.20 | $0.006 | 112 | 0.69 s | 100% | Oct 23 |
| Novita AIfp8 · 1M ctx · 384K out | $0.30 | $1.20 | $0.006 | — | — | — | Oct 23 |
| Nebiusfp8 · 1M ctx | $0.30 | $1.20 | $0.006 | 156 | 1.22 s | 94.9% | Oct 23 |
| DigitalOcean1M ctx | $0.30 | $1.20 | $0.006 | 197 | 0.73 s | 100% | Oct 23 |
| InferenceNetfp8 · 1.04M ctx · 384K out | $0.30 | $1.00 | $0.007 | 94 | 0.40 s | 100% | Oct 23 |
| Crusoefp8 · 1M ctx · 375K out | $0.29 | $1.20 | $0.007 | 117 | 0.61 s | 100% | Oct 23 |
| BaseTen1M ctx · 256K out | $0.30 | $1.20 | $0.007 | 250 | 0.34 s | 100% | Oct 23 |
| Relaceprecision not stated · 1M ctx | $0.07 | $0.50 | $0.008 | 124 | 0.97 s | 100% | Oct 23 |
| Morphfp8 · 1M ctx | $0.25 | $0.60 | $0.031 | 162 | 0.55 s | 99.7% | Oct 23 |
| Alibabaprecision not read: host terms · 1M ctx · 384K outpeak hours | $0.33 | $1.32 | $0.033 | 73 | 1.04 s | 98.1% | Oct 23 |
10 of 22 providers · prices per million tokens, read Oct 23, 06:00 UTC · speed: OpenRouter, median of the last 30 min, read Oct 23, 06:00 UTC · uptime: OpenRouter, last 24 h · −N%: a discount off the host's list price · under each host: weights precision, context and max output as served, the maker's value in brackets when the host's differs
| Provider | Weights | Context | Max output | Tools | Structured outputs | Reasoning |
|---|---|---|---|---|---|---|
| DeepInfra | fp8 | 1M | 128K | yes | yes | yes |
| Relace | not stated by the host | 1M | not stated by the host | yes | yes | yes |
| Cloudflare | not checked yet | 1M | 375K | yes | yes | yes |
| DeepSeek | not checked yet | 1M | 384K | — | — | — |
| Reka | not read: host terms | 1M | 384K | yes | yes | yes |
| Morph | fp8 | 1M | not stated by the host | yes | yes | yes |
| GMICloud | fp8 | 1.05M | 375K | yes | yes | yes |
| AtlasCloud | fp8 | 1M | 384K | yes | yes | yes |
| Modal | not checked yet | 1M | not checked yet | yes | yes | yes |
| InferenceNet | fp8 | 1.04M | 384K | yes | yes | yes |
| Baidu | fp8 | 1M | 384K | yes | yes | yes |
| Crusoe | fp8 | 1M | 375K | yes | yes | yes |
| Phala | not read: host terms | 1M | 384K | yes | yes | yes |
| SiliconFlow | fp8 | 1M | 384K | yes | yes | yes |
| Together AI | fp8 | 1M | not stated by the host | yes | yes | yes |
| Fireworks AI | not checked yet | 1M | 375K | yes | yes | yes |
| BaseTen | not checked yet | 1M | 256K | yes | no | yes |
| Parasail | fp8 | 1M | 375K | yes | yes | yes |
| Novita AI | fp8 | 1M | 384K | yes | no | yes |
| Nebius | fp8 | 1M | not stated by the host | yes | yes | yes |
| DigitalOcean | not checked yet | 1M | not checked yet | yes | yes | yes |
| Alibaba | not read: host terms | 1M | 384K | yes | no | yes |
Weights, context and max output: read at each host's own catalogue or docs where its terms allow, otherwise as the host declares to OpenRouter; hover a cell for its source and date · tools, structured outputs, reasoning: as declared to OpenRouter · not verified by us
DeepSeek V4.1 Flash 1017 spec sheet, field by field, with sources
DeepSeek V4.1 Flash 1017 spec sheetmaker-declared
What DeepSeek declares about the model itself — read from its own weights repository and documentation, not from any host. What each endpoint actually serves is in the pricing above: the two can differ, and both are true.
- Context window
- to verify
- Parameters
- to verify
- Active experts
- to verify
- Weight precision
- to verify
- License
- MIT — open weights, permissiveScope: the licence on the published weights, and nothing else. It is not the maker's acceptable-use policy, which is a separate document, and it is not the contract you sign with whichever host you call — hosts set their own terms, which this index does not read.
- Knowledge cutoff
- not published by the makermaker's docs · read 2026-10-21
- Modalities
- to verify
- Max output (official)
- 384,000 tokensmaker's docs · read 2026-10-21
- Release date
- October 17, 2026maker's docs · read 2026-10-21
- Maker lifecycle status
- General availabilitymaker's docs · read 2026-10-21
huggingface.co · maker's documentation · read 2026-10-21 · 4 of 10 fields published by the maker
| Date | Provider | Input | Output | Change |
|---|---|---|---|---|
| Oct 23 | Morph | $0.25 → $0.25 | $0.60 → $0.60 | 0% |
Every published change of input or output price, newest first · change on input + output · full series: pricing-history.csv.
| Provider | Chatbot 1,000 in / 500 out | RAG / long context 8,000 in / 500 out | Batch summarize 4,000 in / 1,000 out |
|---|---|---|---|
| DeepInfra | $0.24 | $0.93 | $0.69 |
| Relace | $0.32 | $0.81 | $0.78 |
| Cloudflare | $0.40 | $1.10 | $1.00 |
| DeepSeek | $0.45 | $1.50 | $1.20 |
| Reka | $0.45 | $1.50 | $1.20 |
| Morph | $0.55 | $2.30 | $1.60 |
| GMICloud | $0.54 | $1.80 | $1.44 |
| AtlasCloud | $0.58 | $1.95 | $1.56 |
| Modal | $0.75 | $2.50 | $2.00 |
| InferenceNet | $0.80 | $2.90 | $2.20 |
| Baidu | $0.85 | $2.81 | $2.25 |
| Crusoe | $0.89 | $2.92 | $2.36 |
| Phala | $0.90 | $3.00 | $2.40 |
| SiliconFlow | $0.90 | $3.00 | $2.40 |
| Together AI | $0.90 | $3.00 | $2.40 |
| Fireworks AI | $0.90 | $3.00 | $2.40 |
| BaseTen | $0.90 | $3.00 | $2.40 |
| Parasail | $0.90 | $3.00 | $2.40 |
| Novita AI | $0.90 | $3.00 | $2.40 |
| Nebius | $0.90 | $3.00 | $2.40 |
| DigitalOcean | $0.90 | $3.00 | $2.40 |
| Alibaba | $0.99 | $3.30 | $2.64 |
What each host bills for 1,000 requests of a representative job
Your own workload
Other DeepSeek models · all its models, hosts and history ›
- DeepSeek V4.1 Flash29 providers · from $0.47
- DeepSeek V4 Flash 073125 providers · from $0.21
- DeepSeek V4 Pro 081323 providers · from $2.50
- DeepSeek V4 Pro 042315 providers · from $2.40
- DeepSeek V3.214 providers · from $0.52
- DeepSeek V4 Flash 042312 providers · from $0.21