DeepSeek V4.1 Flash 1017 is a DeepSeek model with 1M context, $0.30 in / $1.20 out per million tokens, last verified October 25, 2026. Open weights: yes. Public-board scores are linked to the publisher; vendor-only figures stay in the watch-out, not on the board.
DeepSeek’s October 17 MIT-licensed text/image checkpoint update, currently served through the rolling deepseek-flash API.
Recorded profile verification October 25, 2026·Individual provider fields may change
Pricing tiers: Current rolling API version is DeepSeek-V4.1-Flash-1017. USD/MTok peak input/output $0.30/$1.20, off-peak $0.15/$0.60; cached input $0.006 peak or $0.003 off-peak. Peak hours: weekdays 01:00–04:00 and 06:00–10:00 UTC excluding Chinese public holidays; all remaining hours, weekends and holidays are off-peak. Search and Fetch each cost $0.75/1,000 successful requests. Server-side calls through DeepSeek Harness, Claude Code or Codex are free until October 31, 2026 at 16:00 UTC (November 1 at 00:00 Beijing time), then billed; standalone API prices are separate. This is the live hosted API quote, not a self-hosting rate; the current hosted context/output limits are 1M/384K.
Best for
Evaluating the current DeepSeek Flash API
Open-weight multimodal agents
Self-hosted research with the exact 1017 weights
Watch out
Native config text_config.max_position_embeddings is 1,048,576; hosted API docs state 1M context and maximum 384K output. Recommended native max_tokens >=256K is a sampling recommendation, not a hard native output cap. Numeric parameter counts and a full native effort/default contract are not separately verified for this exact revision. API low/high/max and thinking/non-thinking remain supported under the unchanged calling contract. Vendor DeepSWE 77.0 uses Mini-SWE-V2 and Terminal-Bench 4.0 uses Claude Code; neither is imported as an independent ranked observation. No September benchmark score is inherited.
History
DeepSeek V4.1 Flash 1017 timeline
Release and status events for this model that passed two-source verification — dated, sourced, and linked.
Answered from the verified figures on this page rather than general guidance.
How much does DeepSeek V4.1 Flash 1017 cost per million tokens?
DeepSeek V4.1 Flash 1017 lists base rates of $0.30 per million input tokens and $1.20 per million output tokens. Estimate both input and output token volume and any additional charges for your workload. Current rolling API version is DeepSeek-V4.1-Flash-1017. USD/MTok peak input/output $0.30/$1.20, off-peak $0.15/$0.60; cached input $0.006 peak or $0.003 off-peak. Peak hours: weekdays 01:00–04:00 and 06:00–10:00 UTC excluding Chinese public holidays; all remaining hours, weekends and holidays are off-peak. Search and Fetch each cost $0.75/1,000 successful requests. Server-side calls through DeepSeek Harness, Claude Code or Codex are free until October 31, 2026 at 16:00 UTC (November 1 at 00:00 Beijing time), then billed; standalone API prices are separate. This is the live hosted API quote, not a self-hosting rate; the current hosted context/output limits are 1M/384K.
What is DeepSeek V4.1 Flash 1017's context window?
DeepSeek V4.1 Flash 1017 accepts about 1M tokens of context. That only matters if you routinely send very long documents, large codebases, or multi-turn histories that approach that limit.
What is DeepSeek V4.1 Flash 1017 best for?
DeepSeek V4.1 Flash 1017 is a balanced tier from DeepSeek. It suits evaluating the current deepseek flash api, open-weight multimodal agents, self-hosted research with the exact 1017 weights. Native config text_config.max_position_embeddings is 1,048,576; hosted API docs state 1M context and maximum 384K output. Recommended native max_tokens >=256K is a sampling recommendation, not a hard native output cap. Numeric parameter counts and a full native effort/default contract are not separately verified for this exact revision. API low/high/max and thinking/non-thinking remain supported under the unchanged calling contract. Vendor DeepSWE 77.0 uses Mini-SWE-V2 and Terminal-Bench 4.0 uses Claude Code; neither is imported as an independent ranked observation. No September benchmark score is inherited.
Can I self-host DeepSeek V4.1 Flash 1017?
DeepSeek V4.1 Flash 1017 publishes open weights, but self-hosting depends on the licence, hardware footprint, quantisation quality, and serving stack. A hosted API is often cheaper until you have measured throughput and concurrency on your own hardware.
