AI Agent StoreDeepSeek · Catalog verified
DeepSeek V4.1 Flash 1017
The October 17 update is available for coding, visual questions and work with tools. DeepSeek reports faster, shorter answers, but independent results for this exact revision are still pending.
Released 2026-10-17 · Catalog added 2026-10-17 UTC · Checked 2026-11-03. Catalog addition and provider launch are separate dates.
We still need comparable independent tests. You can compare the confirmed features and limits below, but we cannot yet give this version an Agent Fit or Practical Value rating. Test it on your own tasks before switching; results for another version or thinking setting may not apply.
What you can use today
Token prices measure the small pieces of text the model reads or produces. Capacity is the amount the service accepts in a request; check the limits below for any fallback to another model. A dash means the rate is unverified.
Exact catalog ID: deepseek/deepseek-v4.1-flash-1017. The model accepts text and images, supports tools and can turn thinking on or off. DeepSeek’s own tests use specific tools and settings, so they do not establish the results you will get in your application.
Available thinking settings: none, low, high, max. We do not yet have comparable independent results for these settings. The documented OpenRouter offer default is high when effort is omitted. A default is not a recommendation for every task.
DeepSeek’s deepseek-flash name now selects this October 17 revision. The retired deepseek-v4-flash and deepseek-v4-flash-vision-exp names also run it. To compare versions, use a pinned hosting offer or downloaded checkpoint rather than those moving names.
OpenRouter lists 1,048,576 context tokens and 393,216 maximum output tokens. DeepSeek’s native API and Gateway instead document 1M context and 384K output; use the limits of the service you call.
The displayed OpenRouter quote is off-peak; its catalog also lists peak rates. Buying directly from DeepSeek costs $0.15 input, $0.003 cached input and $0.60 output per million tokens off-peak; peak prices are double. Peak hours are weekdays 01:00–04:00 and 06:00–10:00 UTC, excluding Chinese public holidays. DeepSeek’s Search and Fetch each cost $0.75 per 1,000 successful requests; the free period for built-in calls through DeepSeek Harness, Claude Code and Codex ends October 31 at 16:00 UTC. Allow for retries and any tool fees. Models that think before answering may also charge for that work.
Provider release and specifications ↗
Inspect the exact catalog model ↗
Test it before switching
Give each model the same examples, tools and definition of success. Record correct outcomes, response time, retries and total cost. A cheaper token price does not guarantee a cheaper completed task.
Compare caching and long-context costs ↗