DeepSeek V4 Pro 0813
DeepSeekReleased Apr 24, 2026Open weights · MITdeepseek-v4
- Intelligence (Epoch)
- 156 · 26th 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-pro
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, $3.90 blended. Weights and context per host
Intelligence for the price
2 models score higher for the same price or less; it ranks 3 of the 58 models at its price or below.
Epoch Capabilities Index against the lowest input + output price per M, log scale · prices Oct 23. Score: Epoch AI, CC BY 4.0, 156 with an interval of 154 to 158.
Against its rivals
Its previous version, the best of its price range and the top score
- DeepSeek V4 Pro 0813this model$2.50156
- Claude Opus 5.5top score$24.00169
- Gemini 3.7 Flashsame price range$4.50158
- Gemini 3.8 Flashsame price range$4.50157
- DeepSeek V3.2previous version$0.52146
Same level for 13× less: GLM-5.3-Flash at StreamLake, $0.19 per M input + output · their Epoch intervals overlap (153 against 156).
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 Pro 0813 at each host, fastest first (the 10 fastest of 17)
- CoreWeave139 tok/s0.49 s to first token
- Together AI132 tok/s0.76 s to first token
- Relace111 tok/s1.12 s to first token
- DekaLLM106 tok/s1.13 s to first token
- Azure OpenAI96 tok/s1.98 s to first token
- DeepInfra81 tok/s1.32 s to first token
- Venice73 tok/s1.08 s to first token
- GMICloud71 tok/s2.86 s to first token
- Alibaba61 tok/s1.56 s to first token
- DigitalOcean54 tok/s0.61 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
- 1599B
- 49B active per token (MoE)
- Memory needed
- 959 GB
- in INT4 · 1919 GB in INT8
| Quantization | Memory | Smallest cloud GPU | Quality |
|---|---|---|---|
| INT4 (Q4) | 959 GB | multi-node, over 640 GB | smallest footprint, minor quality loss |
| INT8 (Q8) | 1919 GB | multi-node, over 640 GB | near-lossless |
| FP16 (full) | 3838 GB | multi-node, over 640 GB | reference quality |
At home: 959 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 |
|---|---|---|---|---|---|---|---|
| DeepSeekprecision not stated · 1M ctx · 384K outoff-peak−50% | $0.66 | $1.98 | $0.022 | — | — | — | Oct 22 |
| StreamLakeprecision not read: host terms · 1000K ctx · 375K out−50%reprices often | $0.66 | $1.98 | $0.022 | 48 | 2.17 s | 100% | Oct 23 |
| NextBitprecision not read: host terms · 1M ctx · 375K out | $1.06 | $3.17 | $0.035 | — | — | 100% | Oct 6 |
| GMICloudfp8 · 1.05M ctx−20% | $1.06 | $3.17 | $0.0352 | 71 | 2.86 s | 100% | Oct 23 |
| Baidufp8 · 1M ctx · 384K outreprices often | $1.32 | $3.96 | $0.042 | 37 | 0.90 s | 100% | Oct 23 |
| CoreWeavefp8 · 1M ctx | $1.31 | $3.96 | $0.044 | 139 | 0.49 s | 99.0% | Oct 23 |
| Novita AIfp8 · 1M ctx · 384K out | $1.32 | $3.96 | $0.044 | — | — | — | Oct 23 |
| Parasailfp8 · 1M ctx | $1.32 | $3.96 | $0.044 | 35 | 0.83 s | 98.7% | Oct 23 |
| DigitalOcean1M ctx | $1.32 | $3.96 | $0.044 | 54 | 0.61 s | 94.6% | Oct 23 |
| SiliconFlowfp8 · 1M ctx · 384K out | $1.32 | $3.96 | $0.044 | 32 | 4.38 s | 100% | Oct 23 |
| Cloudflareprecision not stated · 1M ctx | $1.32 | $3.96 | $0.044 | 43 | 1.92 s | 98.9% | Oct 23 |
| Relacefp4 (fp8) · 1M ctx · 384K outreprices often | $0.50 | $2.00 | $0.05 | 111 | 1.12 s | 99.8% | Oct 23 |
| Rekaprecision not read: host terms · 1M ctx | $1.20 | $3.20 | $0.06 | — | — | 100% | Oct 13 |
| DeepInfrafp8 · 1M ctx · 16K out | $1.30 | $2.60 | $0.10 | 81 | 1.32 s | 99.9% | Oct 23 |
| Alibabaprecision not read: host terms · 1M ctx · 384K out | $1.12 | $3.37 | $0.11 | 61 | 1.56 s | 100% | Oct 23 |
| Together AInvfp4 (fp8) · 1M ctx | $1.32 | $3.96 | $0.13 | 132 | 0.76 s | 98.6% | Oct 23 |
| BaseTen1M ctx · 256K out | $1.32 | $3.96 | $0.13 | — | — | 99.1% | Sep 25 |
| AtlasCloudfp8 · 1M ctx · 384K out | $1.32 | $3.96 | $0.13 | 46 | 1.38 s | 100% | Oct 23 |
| Phalaprecision not read: host terms · 1M ctx · 384K out | $1.45 | $4.36 | $0.15 | 52 | 1.49 s | 100% | Oct 23 |
| Azure OpenAI1M ctx · 375K out | $1.91 | $3.83 | $0.16 | 96 | 1.98 s | 100% | Oct 23 |
| Veniceprecision not read: host terms · 1M ctx · 32K out | $1.65 | $4.95 | $0.17 | 73 | 1.08 s | 100% | Oct 23 |
| DekaLLM1M ctx | $1.08 | $3.80 | $0.18 | 106 | 1.13 s | 99.4% | Oct 23 |
| Nebiusfp8 · 979K ctx (1M) | $1.32 | $3.96 | — | — | — | 99.8% | Oct 8 |
10 of 23 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 | 384K | yes | no | yes |
| DeepSeek | not stated by the host | 1M | 384K | — | — | — |
| StreamLake | not read: host terms | 1000K | 375K | yes | no | yes |
| DeepInfra | fp8 | 1M | 16K | yes | yes | yes |
| GMICloud | fp8 | 1.05M | not read: host terms | yes | no | yes |
| NextBit | not read: host terms | 1M | 375K | — | — | — |
| Reka | not read: host terms | 1M | not read: host terms | yes | yes | yes |
| Alibaba | not read: host terms | 1M | 384K | yes | yes | yes |
| DekaLLM | not checked yet | 1M | not checked yet | yes | yes | yes |
| CoreWeave | fp8 | 1M | not checked yet | yes | yes | yes |
| Novita AI | fp8 | 1M | 384K | yes | no | yes |
| BaseTen | not checked yet | 1M | 256K | — | — | — |
| Parasail | fp8 | 1M | not read: host terms | yes | yes | yes |
| Together AI | nvfp4 | 1M | not stated by the host | yes | yes | yes |
| DigitalOcean | not checked yet | 1M | not checked yet | yes | yes | yes |
| SiliconFlow | fp8 | 1M | 384K | yes | no | yes |
| Cloudflare | not stated by the host | 1M | not stated by the host | yes | yes | yes |
| Nebius | fp8 | 979K | not stated by the host | — | — | — |
| AtlasCloud | fp8 | 1M | 384K | yes | yes | yes |
| Baidu | fp8 | 1M | 384K | yes | yes | yes |
| Azure OpenAI | not checked yet | 1M | 375K | yes | no | yes |
| Phala | not read: host terms | 1M | 384K | yes | no | yes |
| Venice | not read: host terms | 1M | 32K | 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 Pro 0813 spec sheet, field by field, with sources
DeepSeek V4 Pro 0813 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
- 1,650,497,936,906≈ 1.65T
- Active experts
- 6 of 384 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≈ August 13, 2026 third-party catalogue — not the maker's figure
- Maker lifecycle status
- Deprecated — retirement announced for September 14, 2026, still served until then · superseded by DeepSeek V4.1 Flashmaker's docs · read 2026-09-10
Also declared in the weights
- Architecture class
- DeepseekV4ForCausalLM
- Layers
- 61
- Vocabulary
- 129,280 tokens
- Weights first published
- August 13, 2026
- Weights last updated
- August 13, 2026
- Paper
- arXiv:2606.19348
deepseek-ai/DeepSeek-V4-Pro-0813 · config.json @ 72e1d32 · read 2026-10-21 · 7 of 10 fields published by the maker
| Date | Provider | Input | Output | Change |
|---|---|---|---|---|
| Oct 23 | Baidu | $0.15 → $1.32 | $0.44 → $3.96 | +809% |
| Oct 23 | DekaLLM | $0.88 → $1.08 | $3.80 → $3.80 | +4% |
| Oct 23 | Phala | $0.96 → $1.45 | $2.88 → $4.36 | +52% |
| Oct 22 | AtlasCloud | $1.68 → $1.32 | $3.38 → $3.96 | +4% |
| Oct 22 | Baidu | $1.69 → $0.15 | $3.38 → $0.44 | −89% |
| Oct 22 | StreamLake | $0.96 → $0.66 | $1.91 → $1.98 | −8% |
| Oct 22 | Venice | $1.65 → $1.65 | $3.30 → $4.95 | +33% |
| Oct 19 | Baidu | $0.18 → $1.69 | $0.35 → $3.38 | +852% |
| Oct 19 | StreamLake | $0.21 → $0.96 | $0.42 → $1.91 | +358% |
| Oct 17 | Baidu | $1.69 → $0.18 | $3.38 → $0.35 | −90% |
| Oct 17 | StreamLake | $0.96 → $0.21 | $1.91 → $0.42 | −78% |
| Oct 14 | Relace | $0.21 → $0.50 | $8.00 → $2.00 | −70% |
| Oct 13 | Reka | $1.12 → $1.20 | $3.37 → $3.20 | −2% |
| Oct 12 | Baidu | $0.17 → $1.69 | $0.34 → $3.38 | +890% |
| Oct 12 | StreamLake | $0.21 → $0.96 | $0.42 → $1.91 | +358% |
| Oct 11 | Baidu | $1.69 → $0.17 | $3.38 → $0.34 | −90% |
| Oct 11 | StreamLake | $0.95 → $0.21 | $1.90 → $0.42 | −78% |
| Oct 10 | StreamLake | $0.96 → $0.95 | $1.91 → $1.90 | −1% |
| Oct 8 | StreamLake | $0.21 → $0.96 | $0.42 → $1.91 | +358% |
| Oct 3 | Baidu | $0.18 → $1.69 | $0.35 → $3.38 | +852% |
| Oct 2 | Baidu | $0.25 → $0.18 | $0.49 → $0.35 | −28% |
| Oct 2 | StreamLake | $0.24 → $0.21 | $0.49 → $0.42 | −15% |
| Oct 1 | Baidu | $1.69 → $0.25 | $3.38 → $0.49 | −85% |
| Oct 1 | StreamLake | $0.96 → $0.24 | $1.91 → $0.49 | −74% |
| Sep 28 | Baidu | $0.35 → $1.69 | $0.69 → $3.38 | +388% |
| Sep 28 | StreamLake | $0.35 → $0.96 | $0.70 → $1.91 | +175% |
| Sep 27 | Baidu | $0.43 → $0.35 | $0.86 → $0.69 | −19% |
| Sep 27 | StreamLake | $0.42 → $0.35 | $0.84 → $0.70 | −17% |
| Sep 26 | Baidu | $0.79 → $0.43 | $1.58 → $0.86 | −46% |
| Sep 26 | StreamLake | $0.79 → $0.42 | $1.59 → $0.84 | −47% |
| Sep 25 | Baidu | $1.69 → $0.79 | $3.38 → $1.58 | −53% |
| Sep 25 | StreamLake | $0.94 → $0.79 | $1.88 → $1.59 | −16% |
| Sep 24 | NextBit | $1.12 → $1.06 | $3.37 → $3.17 | −6% |
| Sep 24 | StreamLake | $0.96 → $0.94 | $1.91 → $1.88 | −2% |
| Sep 22 | Alibaba | $0.58 → $1.12 | $1.74 → $3.37 | +93% |
| Sep 21 | Baidu | $0.43 → $1.69 | $0.86 → $3.38 | +292% |
| Sep 21 | StreamLake | $0.42 → $0.96 | $0.84 → $1.91 | +126% |
| Sep 20 | Baidu | $0.58 → $0.43 | $1.17 → $0.86 | −26% |
| Sep 20 | StreamLake | $0.57 → $0.42 | $1.15 → $0.84 | −26% |
| Sep 19 | Baidu | $0.77 → $0.58 | $1.53 → $1.17 | −24% |
| Sep 19 | StreamLake | $0.76 → $0.57 | $1.51 → $1.15 | −24% |
| Sep 18 | Alibaba | $1.12 → $0.58 | $3.37 → $1.74 | −48% |
| Sep 18 | Baidu | $1.69 → $0.77 | $3.38 → $1.53 | −55% |
| Sep 18 | StreamLake | $0.96 → $0.76 | $1.91 → $1.51 | −21% |
| Sep 16 | Baidu | $0.95 → $1.69 | $1.90 → $3.38 | +78% |
| Sep 16 | StreamLake | $0.95 → $0.96 | $1.89 → $1.91 | +1% |
| Sep 15 | Baidu | $1.69 → $0.95 | $3.38 → $1.90 | −44% |
| Sep 15 | StreamLake | $0.96 → $0.95 | $1.91 → $1.89 | −1% |
| Sep 14 | Baidu | $0.64 → $1.69 | $1.27 → $3.38 | +166% |
| Sep 14 | StreamLake | $0.63 → $0.96 | $1.25 → $1.91 | +52% |
| Sep 13 | Baidu | $0.83 → $0.64 | $1.66 → $1.27 | −23% |
| Sep 13 | StreamLake | $0.96 → $0.63 | $1.91 → $1.25 | −34% |
| Sep 12 | Baidu | $1.69 → $0.83 | $3.38 → $1.66 | −51% |
| Sep 8 | GMICloud | $1.12 → $1.06 | $3.37 → $3.17 | −6% |
| Sep 7 | Baidu | $0.75 → $1.69 | $1.49 → $3.38 | +126% |
| Sep 7 | Novita AI | $1.11 → $1.32 | $3.33 → $3.96 | +19% |
| Sep 6 | Baidu | $0.93 → $0.75 | $1.86 → $1.49 | −20% |
| Sep 5 | Alibaba | $0.58 → $1.12 | $1.74 → $3.37 | +93% |
| Sep 5 | Baidu | $1.03 → $0.93 | $2.07 → $1.86 | −10% |
| Sep 5 | Novita AI | $1.32 → $1.11 | $3.96 → $3.33 | −16% |
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 | $1.50 | $5.00 | $4.00 |
| DeepSeek | $1.65 | $6.27 | $4.62 |
| StreamLake | $1.65 | $6.27 | $4.62 |
| DeepInfra | $2.60 | $11.70 | $7.80 |
| GMICloud | $2.64 | $10.03 | $7.39 |
| NextBit | $2.64 | $10.03 | $7.39 |
| Reka | $2.80 | $11.20 | $8.00 |
| Alibaba | $2.81 | $10.66 | $7.85 |
| DekaLLM | $2.98 | $10.54 | $8.12 |
| CoreWeave | $3.29 | $12.46 | $9.20 |
| Novita AI | $3.30 | $12.54 | $9.24 |
| BaseTen | $3.30 | $12.54 | $9.24 |
| Parasail | $3.30 | $12.54 | $9.24 |
| Together AI | $3.30 | $12.54 | $9.24 |
| DigitalOcean | $3.30 | $12.54 | $9.24 |
| SiliconFlow | $3.30 | $12.54 | $9.24 |
| Cloudflare | $3.30 | $12.54 | $9.24 |
| Nebius | $3.30 | $12.54 | $9.24 |
| AtlasCloud | $3.30 | $12.54 | $9.24 |
| Baidu | $3.30 | $12.54 | $9.24 |
| Azure OpenAI | $3.82 | $17.19 | $11.47 |
| Phala | $3.63 | $13.78 | $10.16 |
| Venice | $4.12 | $15.68 | $11.55 |
What each host bills for 1,000 requests of a representative job
Your own workload
Also serving DeepSeek V4 Pro 0813, without a published price
No price we can verify, so no rank.
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 V4.1 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
Head-to-head: DeepSeek V4 Pro 0813 (DeepSeek) vs GPT-6 Astra (OpenAI) · Kimi K3 (Moonshot AI) vs DeepSeek V4 Pro 0813 (DeepSeek)