DeepSeek V4 Flash 0731
9301°
Run this model
Offers / Ads
DeepSeek | 1,048,576 token context | $0.01/M input, $1.28/M output. DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total. This re-post-trained revision is suited for coding, reasoning, and agent workflows. This is the GA release of DeepSeek V4 Flash.
Guides for this model
Best local LLM by use case (October 2026): the right open model for every job
How much VRAM do I actually need for local AI?
Open weights vs open source vs API-only: what you are actually allowed to do
What is a GGUF file? Quantized models explained (Q4, Q8, fp8, NF4)
Captain's notes
Notes from 2026-10-13
DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek with 13B active parameters drawn from a 284B total pool, built for coding, reasoning, and agent workflows. The GA release substantially improves on its preview, beating DeepSeek V4 Pro on several agentic benchmarks despite activating far fewer parameters. It suits developers who want frontier-class agent and coding performance at a low API price, or who want to self-host a serious MoE model.
Known for
- Strong agentic coding: high scores on DeepSWE, NL2Repo, and terminal benchmarks
- 1M-token context window for long codebases and extended agent sessions
- Speculative decoding built in via DSpark for faster inference on vLLM
- Adjustable reasoning effort: low, high, or max deliberation per call
- Open weights on Hugging Face, runnable on multi-GPU hardware with vLLM
Watch out for
- No standard Jinja chat template; you must use DeepSeek's custom encoding scripts to format prompts correctly.
- 284B total parameters means self-hosting demands serious multi-GPU hardware — this is not a laptop model.
- Output tokens cost 32× more than input tokens; long agentic runs with verbose output add up.
Prompts to try
Autonomous repo task
You are working in a Python repository that has a Flask API and a PostgreSQL backend. The /users endpoint returns all rows from the users table but has no pagination. Add cursor-based pagination using a created_at timestamp cursor. Show the updated route, any new helper functions, and the SQL query. Explain your reasoning before writing code.
Hard debugging puzzle
A Python async function fetches URLs concurrently with asyncio.gather. Under load it occasionally raises RuntimeError: Session is closed. The session is created once at module level. Walk through every plausible cause, rank them by likelihood, and write a corrected implementation that eliminates the race condition.
Multi-step agent plan
You have access to a bash terminal, a Python interpreter, and the ability to read and write files. A CSV file called sales.csv has columns: date, region, product, revenue. Plan and execute the steps to produce a summary report showing total revenue by region and top product per region, then save it as report.md.
Compare with
- DeepSeek V4.1 FlashOpen-weights model1,048,576 token context · $0.04/M input, $0.68/M output
- DeepSeek V4 Flash 0423Open-weights model1,048,576 token context · $0.09/M input, $0.18/M output
- DeepSeek V4 Pro 0813Open-weights model1,048,576 token context · $1.32/M input, $3.96/M output
- DeepSeek V4 Pro 0423Open-weights model1,048,576 token context · $0.96/M input, $1.91/M output
More text models from deepseek
Questions people ask
- Are the weights open and can I run this locally?
- Yes — the weights are on Hugging Face at deepseek-ai/DeepSeek-V4-Flash-0731. Self-hosting requires multi-GPU hardware capable of handling a 284B-parameter MoE model; DeepSeek's own example uses a 4×GB300 node with vLLM.
- What does it cost and how large is the context?
- Via API it is $0.02 per million input tokens and $0.64 per million output tokens, with a context window of 1,048,576 tokens.
- How does this differ from DeepSeek V4 Pro?
- V4 Flash 0731 activates only 13B parameters versus V4 Pro's larger active count, yet it outperforms V4 Pro on every agentic benchmark in DeepSeek's published comparison, making it the better choice for coding-agent workloads at lower inference cost.
Captain's verdictA rare case where the 'flash' tier actually beats the 'pro' on the tasks that matter most — agent coding — and costs a fraction to run.
Model card on Hugging Face
Captain writes these notes from the model's own description and model card. Prices and context sizes are the provider's at the time of writing and can change.
Ranking badge
Updated Oct 22, 2026
Made this model? Show this badge on your page.
Paste it into your Hugging Face card or Civitai description.
Also ranked: #5 Hottest Text / LLM model · #5 Hottest Text / LLM text model
Not this creator?
Get your own badge
DeepSeek V4 Flash 0731 Popularity Trend
Heat score since 31 Jul 2026
Heat score
9,301°
This week
▲
18.4%
9,278° then, 9,301° now
This month
▲
137%
9,185° then, 9,301° now
Past year
n/a
Wasn't released/tracked
7,000°
Floor
The scale runs from 0° to 9,999°. Everything below 7,000° is kept for models no source reports on yet, which carry a stand-in number there so lists stay complete. 7,000° is where measured data begins: a model with real data but almost no downloads sits at the floor, and the measurements fill the degrees above it. The floor says nothing about the model; only the degrees above it do, and they are what separates one model from another.
+1,938°
OpenRouter tokens
28,896,982,056,080 tokens this model processed through OpenRouter in the last 30 days, prompts and answers together. Every 10,000,000 tokens count as one download in the download term, 2,889,698 in all; the last day's tokens join the adoption rate. Source: OpenRouter (openrouter.ai/rankings).
+363°
Adoption rate
About 69,105 downloads a day, measured between two readings of the sources, never guessed from a lifetime average. Speed weighs a quarter of the download term, so a fresh surge shows without letting a one-day spike outrank years of use.
Show detailed breakdown
Every number on this site can be checked. This model's heat score is 9,301°, and here it is built one ingredient at a time, with the running total after each: the floor, the degrees each source adds, the bonus for the kind of model and what the age of its weights takes away. The colors are the bands of the chart above. The glossary at the end defines every term.
- 1Start at the floor.Why 7,000° and not zero: the scale runs from 0° to 9,999°, and the part below 7,000° is reserved for models no source reports on yet, which carry a stand-in number there so the lists stay complete and never outrank a model with data. 7,000° is where measured data begins. A model with real data but almost no downloads sits here; everything that follows is measurement, and the room above the floor holds ten tenfold steps in downloads at 300° each. The floor is the same for every model, so it decides nothing; the degrees above it do.7,000°running total
- 2Add the downloads.2,889,698 download-equivalents in all, from OpenRouter tokens. Ten times the downloads adds the same 300°, so size matters but never runs away with the score.+1,938°8,938°running total
- 3Add the adoption rate.About 69,105 downloads a day right now, measured between two readings, at a quarter of the weight.+363°9,301°running total
- =The heat score.Where this model stands on every list of this site. Nothing else adjusts it.9,301°
Glossary: the terms used here
- Heat score
- One number for how much a model is used and wanted right now, from 0° to 9,999°. Every list on this site that says hottest is sorted by it and nothing else.
- Floor
- The 7,000° every model with real data starts from. The scale runs from 0° to 9,999°; the part below the floor is reserved for stand-ins (models without data), so that a model with data always ranks above one without. The floor is the same for everyone and decides nothing; the degrees above it are the measurement.
- Download-equivalents
- Everything the score treats as a download, added up: downloads themselves, each torrent seeder as 1,000, each AI Horde image and each run on a generation site as one.
- Log scale
- Counting by tens, not by ones: going from 1,000 to 10,000 downloads adds the same degrees as going from 10,000 to 100,000. It keeps giants from crowding out everything else.
- Likes
- Thumbs up and likes from every source, worth half a download each on the log scale.
- Adoption rate
- How many downloads a day a model gets right now, from the difference between two readings of a source. A quarter of the weight of the download term.
- Measured and estimated
- Measured means two readings of a source exist and the difference is real. Estimated means only today's total is known and the past is spread over the model's life; the site says which one it is showing.
- Seeder
- Someone who keeps a copy of the weights online for others to download over a torrent, or a hosted mirror doing the same. Each one counts as 1,000 downloads.
- AI Horde
- A volunteer network that generates images for free on donated computers. The number of images it has made with a model is public, and each counts as one download.
- Generation sites
- Sites that host their own copies of a model and let people generate with it, such as PixAI, SeaArt and Yodayo. Their downloads count like Civitai's and each generation run as one download.
- OpenRouter tokens
- A text model served through an API, such as Claude or GPT, has no weights to download, so its usage is measured in tokens: the pieces of text it read and wrote through OpenRouter, a gateway many apps use, over the last 30 days. Every 10,000,000 tokens count as one download. Source: OpenRouter (openrouter.ai/rankings), CC BY 4.0.
- Full model (checkpoint)
- The weights themselves, the file that generates on its own. It gets a bonus because add-ons depend on it.
- LoRA and embedding
- Small add-on files that steer someone else's weights toward a style, a character or a concept. They cannot run alone.
- Family and weights
- The base a model is built on, such as SD 1.5, SDXL, Flux or Wan. Every model inherits the release date of its family's weights.
- Age
- Degrees taken off for how long ago the family's weights came out: 150° a year, up to 750°. It keeps old downloads that had years to pile up from freezing the top of the list.
- Family boost
- A small, bounded nudge for LoRAs of a family this site is pushing forward right now, always listed in the breakdown.
- Stand-in score
- A stable number below the floor given to a model no source reports on yet, so lists stay complete. It never ranks above a model with data.
- New! (grace period)
- A model less than 7 days old shows New! instead of a score and holds no place in the rankings. Its score is computed from day one, but a model one day old grows faster than anything, so without the wait the week's #1 would always be whatever came out yesterday rather than what held up for a week.
- Trend
- How a model's growth over a period compares with the typical model's growth over the same period. Faster than typical reads as a rise, slower as a fall.
- Typical model
- The middle of the catalog: half the models with a figure grew faster over the period, half slower.
How this model's popularity has moved against the rest of the catalog: its growth in downloads, likes and usage over each period, measured against the typical model's growth over the same period, with its heat score at the time. Above the typical growth reads as a rise, below it as a fall. The past is reconstructed from today's totals, the model's first day and its measured download speed where two readings exist; a model that did not exist at the start of a period shows n/a.
DeepSeek V4 Flash 0731 is a text model for the Text / LLM family listed on ModelTrends.ai, a read-only catalog of open source AI image, video and text models. Compare it with other Text / LLM models and check its heat score to see how it is trending.
Browse by Model Family
Text / LLM Models
434
Anima Models
528
Chroma Models
19
Fibo Models
2
Flux Models
1,496
Flux 2 / Klein Models
1,239
Flux 3 Models
3
Moss Models
3
Iris Models
1
MiniMax H3 Models
364
Hunyuan Models
204
Ideogram Models
20
Krea2 Models
717
Qwen Models
20
Qwen2 Models
47
Ltx2 Models
153