DeepSeek V4 Flash · DeepSeek

DeepSeek V4: Flash vs Pro, and what V4 Flash costs in a coding agent

DeepSeek V4 is DeepSeek's open-weight model line with a 1M-token context, in two sizes: V4 Flash, the small, cheap model, and V4 Pro, the larger model for harder reasoning. DeepSeek has updated both since. V4.1 Flash, released September 10, 2026, replaced V4 Flash, so the old deepseek-v4-flash API name now runs V4.1 Flash at $0.30 / $1.20 per million input and output tokens at peak hours. V4 Pro 0813, from August 13, 2026, is what DeepSeek's API serves as deepseek-v4-pro, at $1.32 / $3.96. Both cost half off-peak. Source: DeepSeek API models and pricing

DeepSeek V4 Flash now bills as DeepSeek V4.1 Flash, which costs about $0.0067 per average coding-agent request on our measured traffic, at list prices of $0.30 per million input and $1.20 per million output tokens. The newer model scores 39 on the Artificial Analysis Intelligence Index. Source: Artificial Analysis (artificialanalysis.ai)

List prices: DeepSeek API pricing, checked 2026-10-13. Request costs: Standard Compute gateway data, 79,301 coding-agent requests from 75 users, Sep 13 to Oct 12, 2026. How we compute this

A newer version is out. DeepSeek V4.1 Flash →

Run DeepSeek V4 Flash in your agent

Or let the router send each step to the cheapest model that can do it, on providers we vet for price, uptime and speed. Plans from $19 a month.

See plans

What DeepSeek V4 Flash costs per coding-agent request

Cost at list price	DeepSeek V4 Flash
Prompt under 100K tokens50% of requests, 87% of input cached	$0.0033
Prompt of 100K to 272K39% of requests, 88% of input cached	$0.0080
Prompt over 272K tokens11% of requests, 86% of input cached	$0.018
Average request	$0.0067
100 requests	$0.67
Median session16 requests; a session ends after 10 idle minutes	$0.091
A month of regular usemedian of 20 users active on 5+ days	$9.13

Each logged request priced at the vendor’s list rates for its own token counts, including cache reads and any higher rate for long prompts. Average output: 698 tokens per request.

DeepSeek V4 Flash benchmarks, speed and latency

Artificial Analysis	Value	Rank
Intelligence Index10 evaluations, 0 to 100	39	18 of 48
Terminal-Bench 4.0agentic coding in a terminal	26.8%	15 of 39
Output speedmedian	195 tokens/s	6 of 44
Time to first answer tokenincludes thinking time	11 s	5 of 44

Source: Artificial Analysis (artificialanalysis.ai). Measured variant: DeepSeek V4.1 Flash (Max), speed on DeepSeek. Fetched October 13, 2026. Rank is among the models on this site, older versions included. Artificial Analysis runs Terminal-Bench 4.0 in its own setup, so its scores differ from the tbench.ai leaderboard.

About DeepSeek V4 Flash

DeepSeek V4 Flash was the small, low-cost member of DeepSeek's V4 line, with open weights and a 1M-token context. DeepSeek has retired the deepseek-v4-flash API name: requests now run on DeepSeek V4.1 Flash and bill at its prices, so the benchmark and cost figures on this page are V4.1 Flash's. Third-party hosts of the original V4 Flash weights set their own prices.

Sources: Artificial Analysis

DeepSeek V4 Flash FAQ

How much does DeepSeek V4 Flash cost per coding-agent request?

$0.0067 on average at list price, $0.67 per 100 requests, and $0.091 for a median 16-request session. That is 79,301 real coding-agent requests (Sep 13 to Oct 12, 2026) priced at DeepSeek's list rates; the median request sent 99,534 input tokens, 87% of input was read from cache, and requests returned 698 output tokens on average.

Does DeepSeek V4 Flash charge more for long prompts?

No. DeepSeek V4 Flash bills the same rate per token up to its 1M-token context window.

How fast is DeepSeek V4 Flash?

Artificial Analysis measures DeepSeek V4.1 Flash (Max) at 195 output tokens per second, with 11 seconds to the first answer token, thinking time included, on DeepSeek's API. Source: Artificial Analysis (artificialanalysis.ai).

Is DeepSeek V4 Flash good for AI agents?

Existing setups that still call the deepseek-v4-flash name. On our scoring, which is calibrated to public benchmarks, it rates 7.1/10 for agentic ability and 7.7/10 for output quality. Open weights.

What is DeepSeek V4 Flash's context window?

DeepSeek V4 Flash has a 1.0M-token context window, large enough to hold sizeable codebases or document sets in a single run. DeepSeek is the maker.

Is DeepSeek V4 Flash open source?

Yes. DeepSeek V4 Flash ships with open weights, so you can self-host it or run it through a hosted provider such as OpenRouter. Check the license before commercial use.

What are the downsides of DeepSeek V4 Flash?

The main trade-offs: Retired as a DeepSeek API name; calls run V4.1 Flash instead; Third-party hosts of the original weights vary in price and speed. It's strongest for existing setups that still call the deepseek-v4-flash name.