DeepSeek V4 Flash 0731
DeepSeekReleased Apr 24, 2026Open weights · MITdeepseek-v4-flash
- Intelligence (Epoch)
- 155 · 29th of 111
- Reasoning
- yes, with effort levels
- Knowledge cutoff
- March 2026
- Context
- 1M tokens
- Max output
- 393k tokens
- Input
- text
- API features
- tools, structured outputs, JSON mode, seed, logprobs
- DeepSeek's API
- yes, as deepseek-v4-flash
Read Oct 23, 06:00 UTC from the hosts' published offers and OpenRouter, spec sheet from DeepSeek · cutoff: aiknowledgecutoff.com
The cheapest host, Relace, serves fp4 weights (maker: fp8). Cheapest at the maker's precision and context: DeepInfra, $0.24 blended. Weights and context per host
Intelligence for the price
On the best-deal line: no model scores higher for less than $0.21 per M.
Epoch Capabilities Index against the lowest input + output price per M, log scale · prices Oct 23. Score: Epoch AI, CC BY 4.0, 155 with an interval of 152 to 157.
Against its rivals
The best of its price range and the top score
- DeepSeek V4 Flash 0731this model$0.21155
- Claude Opus 5.5top score$24.00169
- GLM-5.3-Flashsame price range$0.19153
- DeepSeek V4 Flash 0423same price range$0.21146
Bars: Epoch Capabilities Index (Epoch AI, CC BY 4.0), drawn from 100, below the lowest score in the catalogue · price: lowest input + output per M.
Speed by provider
Output speed of DeepSeek V4 Flash 0731 at each host, fastest first (the 10 fastest of 20)
- BaseTen245 tok/s0.27 s to first token
- Cohere172 tok/s0.39 s to first token
- CoreWeave130 tok/s0.39 s to first token
- DigitalOcean128 tok/s0.73 s to first token
- Parasail84 tok/s0.60 s to first token
- Together AI76 tok/s0.62 s to first token
- GMICloud76 tok/s2.18 s to first token
- Relace75 tok/s0.38 s to first token
- Cloudflare73 tok/s0.72 s to first token
- Venice68 tok/s0.96 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
- 284B
- 13B active per token (MoE)
- Memory needed
- 170 GB
- in INT4 · 341 GB in INT8
- Smallest cloud setup
- 4× H100
- 320 GB · $8.60/h, about $6,278 a month
- Worth it above
- 30B tokens
- a month, against the cheapest API at $0.21 per M
| Quantization | Memory | Smallest cloud GPU | Quality |
|---|---|---|---|
| INT4 (Q4) | 170 GB | 4× H100 · $8.60/h | smallest footprint, minor quality loss |
| INT8 (Q8) | 341 GB | 8× H100 · $17.20/h | near-lossless |
| FP16 (full) | 682 GB | multi-node, over 640 GB | reference quality |
At home: in INT4 it fits a Mac Studio M5 Ultra (256 GB unified), about $9,499 once, or $0.3615/h over 3 years, electricity aside; it pays for itself after about 45B tokens against the cheapest API.
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 |
|---|---|---|---|---|---|---|---|
| StreamLakefp8 · 1000K ctx · 375K out−70%reprices often | $0.13 | $0.40 | $0.0042 | 42 | 1.93 s | 99.9% | Oct 23 |
| DeepSeekprecision not stated · 1M ctx · 375K out | $0.22 | $0.66 | $0.007 | — | — | — | Sep 9 |
| GMICloudfp8 · 1.05M ctx−35% | $0.29 | $0.86 | $0.0091 | 76 | 2.18 s | 100% | Oct 23 |
| Relacefp4 (fp8) · 1M ctxreprices often | $0.07 | $0.14 | $0.014 | 75 | 0.38 s | 100% | Oct 23 |
| Baidufp8 · 1M ctx · 128K outreprices often | $0.44 | $1.32 | $0.014 | 61 | 1.15 s | 100% | Oct 23 |
| Cloudflareprecision not stated · 1M ctx | $0.44 | $1.32 | $0.014 | 73 | 0.72 s | 99.9% | Oct 23 |
| DeepInfrafp8 · 1M ctx · 375K out | $0.06 | $0.18 | $0.015 | 57 | 0.46 s | 100% | Oct 23 |
| AkashMLprecision not stated · 1M ctx | $0.065 | $0.18 | $0.016 | — | — | — | Sep 5 |
| Rekaprecision not read: host terms · 256K ctx (1M) · 128K out | $0.05 | $0.53 | $0.02 | 50 | 0.93 s | 100% | Oct 23 |
| Inceptronfp4 (fp8) · 1M ctxreprices often | $0.11 | $0.41 | $0.027 | 54 | 0.88 s | 99.8% | Oct 23 |
| BaseTenfp8 · 1M ctx · 375K out | $0.13 | $0.26 | $0.028 | 245 | 0.27 s | 99.9% | Oct 23 |
| Veniceprecision not read: host terms · 1M ctx · 32K out | $0.13 | $0.26 | $0.028 | 68 | 0.96 s | 100% | Oct 23 |
| DigitalOcean1M ctx | $0.14 | $0.28 | $0.028 | 128 | 0.73 s | 100% | Oct 23 |
| SiliconFlowfp8 · 1M ctx · 384K out | $0.22 | $0.66 | $0.028 | 25 | 1.49 s | 100% | Oct 23 |
| Phalaprecision not read: host terms · 1M ctx · 384K out | $0.44 | $1.32 | $0.028 | 68 | 2.34 s | 100% | Oct 23 |
| Novita AIfp8 · 1M ctx · 384K out | $0.44 | $1.32 | $0.028 | — | — | — | Oct 23 |
| AtlasCloudfp4 (fp8) · 1M ctx · 384K out | $0.44 | $1.32 | $0.028 | 48 | 1.36 s | 100% | Oct 23 |
| Together AInvfp4 (fp8) · 1M ctx | $0.14 | $0.28 | $0.03 | 76 | 0.62 s | 100% | Oct 23 |
| Azure OpenAI1M ctx · 375K out | $0.21 | $0.56 | $0.031 | 64 | 1.15 s | 95.0% | Oct 23 |
| Alibabaprecision not read: host terms · 1M ctx · 384K out | $0.35 | $1.06 | $0.0352 | 58 | 1.53 s | 98.9% | Oct 23 |
| Morphbf16 (fp8) · 1M ctx | $0.14 | $0.40 | $0.036 | — | — | — | Oct 23 |
| Parasailfp8 · 1M ctx | $0.14 | $0.28 | $0.05 | 84 | 0.60 s | 99.7% | Oct 23 |
| CoreWeavefp8 · 256K ctx (1M) | $0.13 | $0.28 | $0.07 | 130 | 0.39 s | 100% | Oct 23 |
| Cohereprecision not read: host terms · 1M ctx · 375K out | $0.14 | $0.28 | $0.07 | 172 | 0.39 s | 100% | Oct 23 |
| Nebiusfp8 · 1000K ctx | $0.14 | $0.28 | — | — | — | 55.1% | Oct 6 |
10 of 25 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 |
|---|---|---|---|---|---|---|
| Relace | fp4 | 1M | not stated by the host | yes | no | yes |
| DeepInfra | fp8 | 1M | 375K | yes | yes | yes |
| AkashML | not stated by the host | 1M | not stated by the host | — | — | — |
| BaseTen | fp8 | 1M | 375K | yes | no | yes |
| Venice | not read: host terms | 1M | 32K | yes | yes | yes |
| CoreWeave | fp8 | 256K | not checked yet | yes | yes | yes |
| DigitalOcean | not checked yet | 1M | not checked yet | yes | yes | yes |
| Together AI | nvfp4 | 1M | not stated by the host | yes | yes | yes |
| Parasail | fp8 | 1M | not read: host terms | yes | yes | yes |
| Nebius | fp8 | 1000K | not stated by the host | — | — | — |
| Cohere | not read: host terms | 1M | 375K | yes | yes | yes |
| Inceptron | fp4 | 1M | not stated by the host | yes | yes | yes |
| StreamLake | fp8 | 1000K | 375K | yes | no | yes |
| Morph | bf16 | 1M | not stated by the host | — | — | — |
| Reka | not read: host terms | 256K | 128K | yes | yes | yes |
| Azure OpenAI | not checked yet | 1M | 375K | yes | no | yes |
| SiliconFlow | fp8 | 1M | 384K | yes | yes | yes |
| DeepSeek | not stated by the host | 1M | 375K | — | — | — |
| GMICloud | fp8 | 1.05M | not read: host terms | yes | no | yes |
| Alibaba | not read: host terms | 1M | 384K | yes | yes | yes |
| Baidu | fp8 | 1M | 128K | yes | yes | yes |
| Phala | not read: host terms | 1M | 384K | yes | yes | yes |
| Novita AI | fp8 | 1M | 384K | yes | no | yes |
| AtlasCloud | fp4 | 1M | 384K | yes | yes | yes |
| Cloudflare | not stated by the host | 1M | not stated by the host | yes | yes | 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 Flash 0731 spec sheet, field by field, with sources
DeepSeek V4 Flash 0731 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
- 1,048,576 tokens
- Parameters
- 304,180,418,494≈ 304.2B
- Active experts
- 6 of 256 active per token
- Weight precision
- fp8
- License
- MIT — open weights, permissivecurated · read 2026-09-07Scope: 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 maker≈ March 2026 third-party estimate — not the maker's figurecurated
- Modalities
- text
- Max output (official)
- to verify
- Release date
- to verify≈ July 31, 2026 third-party catalogue — not the maker's figure
- Maker lifecycle status
- Retired · superseded by DeepSeek V4.1 Flashmaker's docs · read 2026-09-10
Also declared in the weights
- Architecture class
- DeepseekV4ForCausalLM
- Layers
- 43
- Vocabulary
- 129,280 tokens
- Weights first published
- July 31, 2026
- Weights last updated
- August 1, 2026
- Paper
- arXiv:2606.19348
deepseek-ai/DeepSeek-V4-Flash-0731 · config.json @ 7872f01 · read 2026-10-21 · 7 of 10 fields published by the maker
| Date | Provider | Input | Output | Change |
|---|---|---|---|---|
| Oct 23 | Baidu | $0.0444 → $0.44 | $0.13 → $1.32 | +890% |
| Oct 23 | Inceptron | $0.11 → $0.11 | $0.37 → $0.41 | +10% |
| Oct 23 | Phala | $0.31 → $0.44 | $0.92 → $1.32 | +43% |
| Oct 22 | Baidu | $0.44 → $0.0444 | $1.32 → $0.13 | −90% |
| Oct 22 | Inceptron | $0.13 → $0.11 | $0.45 → $0.37 | −17% |
| Oct 22 | Reka | $0.021 → $0.05 | $0.53 → $0.53 | +5% |
| Oct 22 | StreamLake | $0.0826 → $0.13 | $0.17 → $0.40 | +113% |
| Oct 21 | Inceptron | $0.13 → $0.13 | $0.40 → $0.45 | +9% |
| Oct 19 | Baidu | $0.0598 → $0.44 | $0.18 → $1.32 | +635% |
| Oct 19 | Inceptron | $0.13 → $0.13 | $0.48 → $0.40 | −13% |
| Oct 19 | StreamLake | $0.028 → $0.0826 | $0.056 → $0.17 | +195% |
| Oct 18 | Inceptron | $0.13 → $0.13 | $0.60 → $0.48 | −16% |
| Oct 17 | Baidu | $0.44 → $0.0598 | $1.32 → $0.18 | −86% |
| Oct 17 | Inceptron | $0.14 → $0.13 | $0.60 → $0.60 | −1% |
| Oct 17 | StreamLake | $0.0826 → $0.028 | $0.17 → $0.056 | −66% |
| Oct 16 | Inceptron | $0.13 → $0.14 | $0.60 → $0.60 | +1% |
| Oct 15 | Inceptron | $0.15 → $0.13 | $0.60 → $0.60 | −2% |
| Oct 13 | Inceptron | $0.13 → $0.15 | $0.60 → $0.60 | +2% |
| Oct 12 | Baidu | $0.13 → $0.44 | $0.39 → $1.32 | +239% |
| Oct 12 | Inceptron | $0.17 → $0.13 | $0.60 → $0.60 | −4% |
| Oct 11 | Baidu | $0.44 → $0.13 | $1.32 → $0.39 | −70% |
| Oct 11 | Inceptron | $0.0533 → $0.17 | $0.60 → $0.60 | +17% |
| Oct 10 | Inceptron | $0.05 → $0.0533 | $0.65 → $0.60 | −7% |
| Oct 9 | StreamLake | $0.07 → $0.0826 | $0.14 → $0.17 | +18% |
| Oct 8 | StreamLake | $0.042 → $0.07 | $0.084 → $0.14 | +67% |
| Oct 8 | Venice | $0.17 → $0.13 | $0.35 → $0.26 | −26% |
| Oct 5 | StreamLake | $0.028 → $0.042 | $0.056 → $0.084 | +50% |
| Oct 3 | StreamLake | $0.042 → $0.028 | $0.084 → $0.056 | −33% |
| Oct 1 | AtlasCloud | $0.16 → $0.44 | $0.48 → $1.32 | +178% |
| Oct 1 | Inceptron | $0.056 → $0.05 | $0.65 → $0.65 | −1% |
| Oct 1 | StreamLake | $0.0819 → $0.042 | $0.16 → $0.084 | −49% |
| Sep 30 | Inceptron | $0.056 → $0.056 | $0.36 → $0.65 | +68% |
| Sep 30 | StreamLake | $0.0886 → $0.0819 | $0.18 → $0.16 | −8% |
| Sep 29 | AtlasCloud | $0.44 → $0.16 | $1.32 → $0.48 | −64% |
| Sep 29 | Inceptron | $0.039 → $0.056 | $0.40 → $0.36 | −4% |
| Sep 28 | Baidu | $0.13 → $0.44 | $0.39 → $1.32 | +239% |
| Sep 28 | Inceptron | $0.0828 → $0.039 | $0.41 → $0.40 | −12% |
| Sep 28 | Reka | $0.088 → $0.021 | $0.53 → $0.53 | −11% |
| Sep 28 | StreamLake | $0.047 → $0.0886 | $0.0941 → $0.18 | +88% |
| Sep 27 | Baidu | $0.0598 → $0.13 | $0.18 → $0.39 | +117% |
| Sep 26 | Baidu | $0.44 → $0.0598 | $1.32 → $0.18 | −86% |
| Sep 26 | StreamLake | $0.049 → $0.047 | $0.098 → $0.0941 | −4% |
| Sep 25 | StreamLake | $0.0886 → $0.049 | $0.18 → $0.098 | −45% |
| Sep 24 | Inceptron | $0.06 → $0.0828 | $0.30 → $0.41 | +38% |
| Sep 22 | Alibaba | $0.18 → $0.35 | $0.53 → $1.06 | +100% |
| Sep 22 | Baidu | $0.048 → $0.44 | $0.14 → $1.32 | +817% |
| Sep 22 | Inceptron | $0.08 → $0.06 | $0.20 → $0.30 | +29% |
| Sep 21 | Reka | $0.11 → $0.088 | $0.66 → $0.53 | −20% |
| Sep 21 | StreamLake | $0.0367 → $0.0886 | $0.0734 → $0.18 | +141% |
| Sep 20 | StreamLake | $0.0473 → $0.0367 | $0.0946 → $0.0734 | −22% |
| Sep 19 | Baidu | $0.44 → $0.048 | $1.32 → $0.14 | −89% |
| Sep 19 | StreamLake | $0.0484 → $0.0473 | $0.0969 → $0.0946 | −2% |
| Sep 18 | Alibaba | $0.35 → $0.18 | $1.06 → $0.53 | −50% |
| Sep 18 | Inceptron | $0.0606 → $0.08 | $0.20 → $0.20 | +6% |
| Sep 18 | StreamLake | $0.0886 → $0.0484 | $0.18 → $0.0969 | −45% |
| Sep 17 | Inceptron | $0.0549 → $0.0606 | $0.17 → $0.20 | +15% |
| Sep 16 | StreamLake | $0.0855 → $0.0886 | $0.17 → $0.18 | +4% |
| Sep 15 | Baidu | $0.0352 → $0.44 | $0.11 → $1.32 | +1150% |
| Sep 15 | StreamLake | $0.0886 → $0.0855 | $0.18 → $0.17 | −3% |
| Sep 14 | StreamLake | $0.0498 → $0.0886 | $0.0997 → $0.18 | +78% |
| Sep 13 | Baidu | $0.44 → $0.0352 | $1.32 → $0.11 | −92% |
| Sep 13 | StreamLake | $0.0668 → $0.0498 | $0.13 → $0.0997 | −25% |
| Sep 12 | StreamLake | $0.0886 → $0.0668 | $0.18 → $0.13 | −25% |
| Sep 10 | Baidu | $0.14 → $0.44 | $0.28 → $1.32 | +319% |
| Sep 8 | GMICloud | $0.35 → $0.29 | $1.06 → $0.86 | −19% |
| Sep 7 | Baidu | $0.05 → $0.14 | $0.10 → $0.28 | +180% |
| Sep 7 | DeepInfra | $0.08 → $0.06 | $0.18 → $0.18 | −8% |
| Sep 7 | Novita AI | $0.41 → $0.44 | $1.23 → $1.32 | +8% |
| Sep 5 | Alibaba | $0.18 → $0.35 | $0.53 → $1.06 | +100% |
| Sep 5 | Baidu | $0.065 → $0.05 | $0.13 → $0.10 | −23% |
| Sep 5 | Novita AI | $0.44 → $0.41 | $1.32 → $1.23 | −7% |
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 |
|---|---|---|---|
| Relace | $0.14 | $0.63 | $0.42 |
| DeepInfra | $0.15 | $0.57 | $0.42 |
| AkashML | $0.15 | $0.61 | $0.44 |
| BaseTen | $0.26 | $1.17 | $0.78 |
| Venice | $0.26 | $1.17 | $0.78 |
| CoreWeave | $0.27 | $1.18 | $0.80 |
| DigitalOcean | $0.28 | $1.26 | $0.84 |
| Together AI | $0.28 | $1.26 | $0.84 |
| Parasail | $0.28 | $1.26 | $0.84 |
| Nebius | $0.28 | $1.26 | $0.84 |
| Cohere | $0.28 | $1.26 | $0.84 |
| Inceptron | $0.32 | $1.09 | $0.85 |
| StreamLake | $0.33 | $1.25 | $0.92 |
| Morph | $0.34 | $1.34 | $0.97 |
| Reka | $0.31 | $0.66 | $0.73 |
| Azure OpenAI | $0.49 | $1.96 | $1.40 |
| SiliconFlow | $0.55 | $2.09 | $1.54 |
| DeepSeek | $0.55 | $2.09 | $1.54 |
| GMICloud | $0.72 | $2.72 | $2.00 |
| Alibaba | $0.88 | $3.34 | $2.46 |
| Baidu | $1.10 | $4.18 | $3.08 |
| Phala | $1.10 | $4.18 | $3.08 |
| Novita AI | $1.10 | $4.18 | $3.08 |
| AtlasCloud | $1.10 | $4.18 | $3.08 |
| Cloudflare | $1.10 | $4.18 | $3.08 |
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 Pro 081323 providers · from $2.50
- DeepSeek V4.1 Flash 101722 providers · from $0.39
- DeepSeek V4 Pro 042315 providers · from $2.40
- DeepSeek V3.214 providers · from $0.52
- DeepSeek V4 Flash 042312 providers · from $0.21