DeepSeek

DeepSeek V4.1 Flash

deepseek/deepseek-v4.1-flashHugging Face ↗

DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the first built on the company's Causal Encoder-Decoder (CED) architecture. It activates 8B parameters on input and 16B on...

toolsstructuredreasoningimage in
Data source

Only tracked on Vercel AI Gateway.

DeepSeek V4.1 Flash usage over time

Vercel AI Gateway share of tokens, by day.

View
Tidelines — LLM usage and market share

DeepSeek V4.1 Flash holds 30.72% today, up from 13.82% at the start of the visible window.

Data: Vercel AI Gateway as of 2026-10-09 · Source: tidelines.ai/models/deepseek-deepseek-v4-1-flash

Rank history

Tidelines — LLM usage and market share
Data: Vercel AI Gateway as of 2026-10-09 · Source: tidelines.ai/models/deepseek-deepseek-v4-1-flash

Adoption curve vs the pantheon

DeepSeek V4.1 Flash's share plotted against days since first tracked, overlaid on other current leaders on the same axis. The question every new-model page asks: is this one going to make it?

Tidelines — LLM usage and market share
Data: Vercel AI Gateway as of 2026-10-09 · Source: tidelines.ai/models/deepseek-deepseek-v4-1-flash

Citable stat

As of 2026-10-09, DeepSeek V4.1 Flash holds 30.7% of Vercel AI Gateway token share, up 30.72pp over 30 days. Source: Tidelines (Vercel AI Gateway).

Cost per session

Median dollars spent on one DeepSeek V4.1 Flash session, by coding agent and session length.

Tidelines — LLM usage and market share

Token rate sets the price; session length and reasoning set the bill.

Agent1 turn2–9 turns10–49 turns50+ turns
Claude Code$0.0041$0.011$0.073$0.641
Hermes Agent$0.0012$0.0071$0.061$0.378
Kilo Code$0.0015$0.014$0.057$0.353
Codex$0.0010$0.0075$0.043$0.107

Compare every model's session cost →

Data: OpenRouter session-cost dataset as of 2026-09-27 · Source: tidelines.ai/models/deepseek-deepseek-v4-1-flash

Where to run it

DeepSeek V4.1 Flash pricing by provider

Tidelines — LLM usage and market share

Per-provider pricing, context, and uptime via OpenRouter. Sorted by output price.

DeepSeek V4.1 Flash is served by 31 tracked endpoints on OpenRouter. The cheapest output price is $0.18 per million tokens from Decart, and the most expensive is $2.40 from BaseTen · 1049k. The largest context window offered is 1049k tokens. Prices are list rates and refresh daily.

ProviderInput $/MOutput $/MContextUptime
Decart$0.09$0.181049k100.00%
Sail Research$0.09$0.301049k100.00%
DeepInfra$0.14$0.421049k100.00%
StreamLake$0.15$0.591024k100.00%
Relace$0.05$0.601049k100.00%
InferenceNet$0.08$0.601040k100.00%
DeepSeek$0.15$0.601049k100.00%
CoreWeave$0.20$0.651049k100.00%
DigitalOcean$0.17$0.661049k100.00%
Morph$0.04$0.691049k100.00%
GMICloud$0.18$0.721049k100.00%
AtlasCloud$0.19$0.771049k100.00%
Novita$0.20$0.781049k100.00%
Phala$0.21$0.841049k100.00%
OpenInference$0.04$0.891049k100.00%
Wafer$0.10$0.991049k100.00%
Cloudflare$0.20$1.001049k100.00%
Ionstream$0.27$1.101049k100.00%
Makora$0.27$1.151049k100.00%
Crusoe$0.29$1.201049k100.00%
Baidu$0.30$1.201049k100.00%
BaseTen · 1049k$0.30$1.201049k100.00%
Alibaba$0.30$1.201000k100.00%
Together$0.30$1.201049k100.00%
SiliconFlow$0.30$1.201049k100.00%
Modal$0.30$1.201049k100.00%
Parasail$0.30$1.201049k100.00%
Fireworks · 1049k$0.30$1.201049k100.00%
Venice$0.30$1.201000k100.00%
Fireworks · 1049k$0.45$1.801049k100.00%
BaseTen · 1049k$0.60$2.401049k100.00%