Get started

DeepSeek V4 Pro 0423

Sold by 13 sellers, timed at 59.4 tok/s, scored 20.8 on the Intelligence Index, on 11 Design Arena boards, and with results on 2 openly licensed boards.

retrieved sha256 · 0bb53a77713f

Models/DeepSeek/DeepSeek V4 Pro 0423USD per 1M tokens · as OpenRouter routes it
deepseek/deepseek-v4-proDeepSeek · base listing · DeepSeek tokenizer
OpenRouter list price · the routed rate, as the catalog carried itretrieved 17 Nov 16:29 UTC from openrouter.ai ↗
Input$0.90per 1M tokens
Output$1.50per 1M tokens
Cached input$0.25per 1M tokens
Context1.05M1,048,576 tokens
What Vitna read · 3 providers, each from its own pricing page, kept as dated recordsin / out per 1M tokens
DeepInfraopen weights$1.30 / $2.60confirmed 22 Nov 2026
Parasailopen weights$1.74 / $3.48confirmed 22 Nov 2026
Novita AIopen weights$1.60 / $3.20confirmed 22 Nov 2026

The same weights served by 3 providers, each at its own list price. OpenRouter's rate is its routed rate across whichever of them it uses, and it need not match any one of them.

What the list says · the fields OpenRouter publishes for this id10 fields
AuthorDeepSeek
Listed
Canonical iddeepseek/deepseek-v4-pro-20260423
Context1,048,576 tokens
Max output393,216 tokens
Modalitiestext → text
TokenizerDeepSeek
Reasoningoptional · effort xhigh, high, none, high by default
Moderationnot moderated by OpenRouter
What the weights say · read from the model's own repository on Hugging Faceretrieved from huggingface.co ↗
Licencemit
Parameters1.6T · 1,598,839,674,782 declared
Model classDeepseekV4ForCausalLM
Downloadnot gated
Published

What sellers charge

13 sellers · 2.95x from cheapest to dearest

One tick per seller, on a log scale of input price per million from $0.66079 at Baidu to $1.95 at DeepInfra. 2 sellers undercut the $0.90 listed; 9 are dearer. A faint tick declares no number format. Every seller is named on the endpoint ledger below.

The same sellers grouped by the number format each declares, on the same scale. The fp8 group alone spans 100% of the range, so the cheap end is not cheap for its number format. A declaration is checked by nobody: this groups the rows and certifies none.

Measured speed

59,755 requests · 13 endpoints · 30-minute window
Throughput59.4 tok/sMedian output rate, weighted across the 13 endpoints serving it by the requests each answered.the speed board
First token1.15sMedian wait before the first token, over the same window and weighted the same way.how Vitna reads a figure
Endpoint spread29 to 95.5The slowest and the fastest route serving this model, in tokens per second. Speed is a property of the route, not of the weights.the board, fastest first

OpenRouter's own timing of its routed traffic, weighted by requests, over one 30-minute window read . Every route is on the endpoint ledger below.

Every endpoint

13 priced · 13 timed · cheapest first
  1. Baidubaidu/fp8 · 61% off$0.66079$1.32158$0.05474fp8393k54 tok/s351.08s11,699 req100%
  2. StreamLakestreamlake/fp8 · 62% off$0.662244$1.324488$0.055187fp8384k42 tok/s331.24s20,572 req100%
  3. Relacerelace/fp4 · listed$0.90$1.50$0.20fp4393k94 tok/s52933ms14,416 req99.93%
  4. SiliconFlowsiliconflow/fp8 · listed$0.90$1.70$0.08fp8393k29 tok/s231.85s1,837 req99.72%
  5. DeepInfradeepinfra/fp8/flex$1.04$2.08$0.08fp8393knot matchednonenonenot matched100%
  6. DigitalOceandigitalocean$1.044$2.088$0.2088not declared384k66 tok/s45574ms3,903 req99.87%
  7. Novitanovita/fp8 · 20% off$1.28$2.56$0.108fp8393k48 tok/s212.14s2,955 req100%
  8. DeepInfradeepinfra/fp8$1.30$2.60$0.10fp8393knot matchednonenonenot matched99.49%
  9. Alibabaalibaba/fp8$1.416$2.832$0.118fp8393knot matchednonenonenot matched100%
  10. Alibabaalibaba/fp8/us-east-1$1.56$3.12$0.13fp8393knot matchednonenonenot matched100%
  11. Venicevenice$1.65$3.301$0.33not declared262k38 tok/s21.62.01s99 req81.77%
  12. Azureazure/us$1.91$3.83$0.16not declared384k95.5 tok/s481.15s739 req100%
  13. DeepInfradeepinfra/fp8/priority$1.95$3.90$0.15fp8393knot matchednonenonenot matched93.63%
  14. Alibabanot matchednot matchednot matchednot matchedno offer38 tok/s291.11s186 reqno offer
  15. Alibabanot matchednot matchednot matchednot matchedno offer44 tok/s26841ms1,534 reqno offer
  16. DeepInfranot matchednot matchednot matchednot matchedno offer53 tok/s303s54 reqno offer
  17. DeepInfranot matchednot matchednot matchednot matchedno offer52.5 tok/s281.33s1,567 reqno offer
  18. DeepInfranot matchednot matchednot matchednot matchedno offer66 tok/s34.21.32s194 reqno offer

Each track runs from a route's p90, the rate nine in ten of its requests met or beat, up to its median. Weighted by requests, the figure above reports where the traffic went rather than what one route does: 9 of the 13 timed routes are slower than it, and the middle route runs at 52.5 tok/s. The widest median-to-p90 gap is Novita's, 2.29x.

Alibaba publishes 2 offers and had 2 routes timed; DeepInfra publishes 3 offers and had 3 routes timed. The two documents share no route id, so those offers and routes are printed apart rather than paired.

Uptime is OpenRouter's own figure for each seller over the trailing day, read once a day, and the column prints the latest reading, taken . 15 of the 19 sellers with a reading fell under 99% on at least one of the 23 days held, 23 Oct 2026 to 21 Nov 2026: Venice at 81.77% on 21 Nov 2026, Baidu at 84.65% on 11 Nov 2026, DeepInfra (deepinfra/fp8/flex) at 87.93% on 18 Nov 2026, Alibaba (alibaba/fp8) at 89.94% on 25 Oct 2026, and 11 more. Source: OpenRouter's public endpoints API, one reading a day for each seller: 1,211 sellers carried one on the latest of the 23 days held.

Source: OpenRouter's public endpoints API. 1257 sellers across 359 models when the mirror ran, 186 of them served by more than one. List prices, not what anybody paid; a discount is the seller's own. Read . Speed is OpenRouter's own timing over one 30-minute window, read .

What each seller accepts

21 parameters · 13 sellers
  1. include_reasoningevery seller
  2. max_tokensevery seller
  3. reasoningevery seller
  4. reasoning_effortevery seller
  5. temperatureevery seller
  6. tool_choiceevery seller
  7. toolsevery seller
  8. top_pevery seller
  9. response_format11 of 13not DeepInfra (deepinfra/fp8/flex) and DeepInfra (deepinfra/fp8/priority)
  10. stop10 of 13not Azure, DigitalOcean, and SiliconFlow
  11. frequency_penalty9 of 13
  12. top_k9 of 13
  13. presence_penalty8 of 13
  14. structured_outputs8 of 13
  15. seed6 of 13
  16. repetition_penalty5 of 13
  17. logit_bias4 of 13
  18. min_p4 of 13
  19. logprobs3 of 13only DigitalOcean, Novita, and StreamLake
  20. top_logprobs3 of 13only DigitalOcean, Novita, and StreamLake
  21. max_completion_tokens1 of 13only Azure

8 of these 21 are accepted by every one of the 13 sellers that publish a list; the other 13 are declared by some sellers and not others, so what a request sending one of them gets depends on which seller serves it. Each list is the seller's own declaration in OpenRouter's endpoints API, read , and nothing here tested one.

Price position

342 priced models

Blended $1.05 per million tokens, 3 in to 1 out: cheaper than 142 of the 342 priced models in the catalog. Each bar is one stretch of the catalog on a logarithmic price scale; the mark is this listing.

Measured quality

145 models with an index · 132 on a design arena board

Each mark is one board this model stands on, placed at the share of that board it scores above, and led to its number in the list under the rule. The columns and the ledger below carry the figures.

Intelligence Index20.8Position 38 of 58, above 34% of the models scored.the intelligence leaderboard
Coding Indexnot scoredArtificial Analysis has not scored this model on its Coding Index.the coding leaderboard
Agentic Indexnot scoredArtificial Analysis has not scored this model on its Agentic Index.the agentic leaderboard
  1. Full stack · Agents948 Elowin rate 22.1% · source rank 46the board, 44 ranked
  2. Godot game development · Agents1059 Elowin rate 34.0% · source rank 32the board, 34 ranked
  3. Web apps · Agents1000 Elowin rate 26.4% · source rank 47the board, 47 ranked
  4. 3D scenes · Models1233 Elowin rate 56.8% · source rank 37the board, 113 ranked
  5. ASCII art · Models1132 Elowin rate 46.4% · source rank 48the board, 67 ranked
  6. Code · Models1229 Elowin rate 52.2% · source rank 51the board, 120 ranked
  7. Data visualisation · Models1199 Elowin rate 48.8% · source rank 61the board, 110 ranked
  8. Game development · Models1221 Elowin rate 53.0% · source rank 48the board, 120 ranked
  9. SVG · Models1128 Elowin rate 45.5% · source rank 59the board, 83 ranked
  10. UI components · Models1209 Elowin rate 50.6% · source rank 58the board, 116 ranked
  11. Websites · Models1226 Elowin rate 50.9% · source rank 59the board, 126 ranked
Epoch Capabilities Index · a fit by a third party over every benchmark this model has results onthe whole index, with intervals
Index150.2148.4 to 152.0
Position71 of 258above 73% of the models placed
Results behind it14benchmark results
As Epoch names itDeepSeek-V4-Proreleased 24 Apr 2026

Position 71 of the 258 models Epoch has placed, which is not the population of either board above, on a scale anchored at Claude 3.5 Sonnet 130 and GPT-5 150. Epoch states that absolute values are meaningless by themselves and that comparisons between models are what the index is for, that values are not linearly related to benchmark accuracy, and that a highly specialised model may score low while being capable in its own domain.

Benchmark results

2 of 6 boards · 37 results
Epoch AI Benchmarking Hub · publisher-run, CC-BY-4.08 of 22 tasks

Mean accuracy, 0 to 1, on the task the row names, as Epoch AI runs it and averaged over repeated runs. All eighteen of their tasks, including the held-out FrontierMath tiers, named as they name them. The model id carries the effort setting the run used. Epoch AI Benchmarking Hub names these runs deepseek-v4-pro_max, deepseek-v4-pro_high and deepseek-v4-pro_none, and the effort beside each score says which run it is. A run missing from a task's rows has no published result there, which is not a score of zero. All runs is how many runs the publisher scored on the task, every model's. Read from epoch.ai ↗, which dates its own data . Epoch AI Benchmarking Hub data, by Epoch AI, under CC BY 4.0.

LiveBench · publisher-run, Apache-2.023 of 23 tasks · 71 runs scored on each task · effort unstated

Percent correct on the task the row names, higher is better. All 23 published columns; there is no average to carry, because the site computes its headline in the browser from a weighting this does not reimplement. LiveBench names this run deepseek-v4-pro, with no effort setting written on it. Read from livebench.ai ↗, which dates its own data . LiveBench, Apache-2.0. Results produced by the LiveBench maintainers on questions refreshed monthly.

2 of the 6 openly licensed boards Vitna mirrors carry a result for DeepSeek V4 Pro 0423, each printed as its publisher published it and none added to another. MTEB, HELM, Vectara hallucination leaderboard and Aider polyglot have not published one under a name that matches its listing, and names are matched to the catalog exactly and never guessed. A task that is a link opens its column on /capability, with every run scored on it; an effort that is a link opens its run's row there, where the run is ranked on Vitna's scale. Every board, with its licence and what its publisher measures, is on /benchmarks.

Sources

Source: OpenRouter (openrouter.ai/models), as of 17 Nov 16:29 UTC. sha256 0bb53a77713f over 853,270 bytes.
Catalog: OpenRouter's public model list.
Intelligence, coding and agentic indexes: Artificial Analysis, as published by OpenRouter.
Elo, win rate and rank: Design Arena, as published by OpenRouter.
Source: each model's own Hugging Face repository (huggingface.co), as of 20 Nov 2026. 162 repositories read, 4 did not answer.
Licence, parameter count, model class and gating: each model's own Hugging Face repository, the one OpenRouter names for the listing.
A licence is the tag the uploader put on the repository, read from nothing else and verified by nobody.
A parameter count is the total the repository's safetensors files declare; for a mixture of experts that is every expert, not the share one token runs through.
Source: the Epoch AI Capabilities Index (epoch.ai/eci), as of 20 Nov 2026. 258 models placed over 2917 benchmark results.
Epoch AI Benchmarking Hub data, by Epoch AI, under CC BY 4.0.
Vitna ran none of these evaluations and adjusted none of these figures.