AI Model Release Tracker

DeepSeek-V4-Pro-0813Open Weight

Released

DeepSeek-V4-Pro-0813 is an AI model released by DeepSeek on Thursday, Aug 13 2026, 13 days after DeepSeek-V4-Flash-0731. It is an open-weight model — the trained weights are available to download and run. It is a 1.6T parameterThe learned values that shape how a model responds. B means billion and T means trillion. For mixture-of-experts models, the total count is larger than the number active for each token. Parameter count alone does not measure quality. model with a 1M token context windowHow much information the model can work with in one request, measured in tokens — pieces of words or other data. This includes your question, attachments, conversation and reply. Limits can vary by model variant and provider.. Benchmark results (shown below) cover BullshitBench v2, SWE-Bench Verified, Terminal-Bench 2.1, Toolathlon-Verified, BrowseComp, CyberGym, and 3 more.

API pricing

Off-peak (including weekends and Chinese public holidays)
InputWhat you pay for everything you send the model — your question, plus any documents or earlier conversation you include with it.
$0.66
Cached inputA reduced rate for text you send over and over. If every request starts with the same instructions or the same document, the provider keeps a copy ready and charges less to read it again.
$0.022
OutputWhat you pay for the text the model writes back. It is normally the dearer half: producing an answer costs more than reading one.
$1.98
Peak (01:00–04:00 and 06:00–10:00 UTC, Monday–Friday except Chinese public holidays)
InputWhat you pay for everything you send the model — your question, plus any documents or earlier conversation you include with it.
$1.32
Cached inputA reduced rate for text you send over and over. If every request starts with the same instructions or the same document, the provider keeps a copy ready and charges less to read it again.
$0.044
OutputWhat you pay for the text the model writes back. It is normally the dearer half: producing an answer costs more than reading one.
$3.96
  • Thinking and non-thinking modes use the same published price.
USD per 1M tokensEvery price here is for one million tokens. A token is roughly three-quarters of a word, so a million tokens is about 750,000 words of text.Verified October 19, 2026Read from api-docs.deepseek.com. Click the check to open the source.

Available from

DeepSeek——$0.66$1.981M—
StreamLake——$0.66$1.981M—
Sail Research——$1.35$2.401Mfp4
Sail Researchus—$1.35$2.401Mfp4
DeepInfra——$1.30$2.601Mfp8
Novita——$0.99$2.971Mfp8
GMICloud——$1.056$3.1681Mfp8
Baidu——$1.3187$3.9561Mfp8
CoreWeave——$1.31$3.961Mfp8
Alibabaus-east-1—$1.32$3.961M—
AtlasCloud——$1.32$3.961Mfp8
BaseTen——$1.32$3.961Mfp4
Cloudflare——$1.32$3.961M—
DigitalOcean——$1.32$3.961M—
Nebius——$1.32$3.961Mfp8
Parasail——$1.32$3.961Mfp8
SiliconFlow——$1.32$3.961Mfp8
Together——$1.32$3.961M—
Wafer——$0.285$4.301Mmxfp4
Waferus—$0.111$4.441Mmxfp4
Venice——$1.65$4.951M—
Parasail—fast$1.98$5.941Mfp8
Relace——$0.0999$8.001Mfp4
USD per 1M tokensEvery price here is for one million tokens. A token is roughly three-quarters of a word, so a million tokens is about 750,000 words of text.Checked November 12, 2026Third-party rates from openrouter.ai, not the lab’s own list price.

Benchmarks

Coding

SWE-Bench VerifiedCoding — Real coding tasks pulled from open-source projects — the AI has to find and fix actual bugs. A human-checked version of the original SWE-Bench. Higher is better.via BenchLM
80.6%
#11 of 80

Terminal & CLI

Terminal-Bench 2.1Agentic terminal coding — Can the AI work in a command-line terminal — running commands and finishing technical setup tasks the way a developer would? Higher is better.
87.9%
#9 of 40

Agentic & tool use

Toolathlon-VerifiedPersonal tool use — Tests how well the AI uses everyday personal tools and apps to get things done — a human-checked version of Toolathlon. Higher is better.
74.1%
#3 of 9
BrowseCompWeb browsing — Can the AI browse the web and track down hard-to-find answers? Higher is better.via BenchLM
83.4%
#17 of 35
CyberGymCybersecurity — Tests the AI on cybersecurity challenges — finding and exploiting software weaknesses inside a safe sandbox. Higher is better.
83.3%
#6 of 17
AutomationBenchBusiness workflows — Tests whether the AI can run real multi-step business workflows — the kind of end-to-end office processes companies want to automate — from start to finish. Higher is better.
31.8%
#19 of 32

Reasoning & science

Humanity's Last ExamMultidisciplinary reasoning — Humanity's Last Exam — extremely hard expert questions across many subjects, written so you can't just look up the answer. “No tools” means the AI answers on its own. Higher is better.
42.7%no tools
60%with tools
no tools
#16 of 48
with tools
#13 of 60
ARC-AGI-2Abstract reasoning — Puzzle-style tests of abstract reasoning and pattern-finding — the kind of thing people find easy but AIs often struggle with. Higher is better.via BenchLM
61.3%
#25 of 41

Robustness

BullshitBench v2Nonsense detection — Given a confidently-worded but nonsensical prompt, does the AI spot that it makes no sense and push back — instead of playing along and inventing an answer? The score is how often it clearly called out the nonsense. Higher is better.via BullshitBench
35%
#57 of 100

Source: BenchLM, retrieved 28 September 2026. Other sources are identified on the linked benchmark pages.

Compare DeepSeek-V4-Pro-0813 with

DeepSeek-V4-Pro-0813

Suggested comparisons

DeepSeek-V4-Pro-0813vsDeepSeek-V4.1-Flash-1017DeepSeek-V4-Pro-0813vsGPT-6.1 TerraDeepSeek-V4-Pro-0813vsClaude Fable 5.2DeepSeek-V4-Pro-0813vsGemini 4 CarbonDeepSeek-V4-Pro-0813vsMuse EmberDeepSeek-V4-Pro-0813vsGrok 4.7DeepSeek-V4-Pro-0813vsMistral Large 4DeepSeek-V4-Pro-0813vsKimi K3.1 ProDeepSeek-V4-Pro-0813vsGLM-5.3-FlashDeepSeek-V4-Pro-0813vsQwen4-Max PreviewDeepSeek-V4-Pro-0813vsNemotron 3.5 Bolt E4B 4GB

About

DeepSeek V4 Pro 0813 was introduced in August 2026 as the general-availability update to the Pro tier of the V4 family. DeepSeek emphasized stronger agent performance and made the model available through the app, website and API. In the hosted assistant, the updated Pro model was served through the existing Expert Mode.

The announcement added selectable low, high and maximum reasoning effort for both V4 Pro and V4 Flash. It also extended to the Pro tier the native Responses API and Codex support that V4 Flash 0731 had introduced, while retaining existing API model names. DeepSeek published the 0813 checkpoint’s weights under the MIT License, providing a self-hosted option alongside the hosted service.

Frequently asked questions

DeepSeek-V4-Pro-0813 was released by DeepSeek on Thursday, Aug 13 2026.

All DeepSeek releases

24 tracked