DeepSeek upgrades V4.1 Flash model and opens its search API
In short: DeepSeek released DeepSeek-V4.1-Flash-1017, an upgraded version of its Flash model, accessible through the same API endpoint and pricing as before. The update brings notable gains across agentic, coding, and tool-use benchmarks while reducing task runtime and output length. DeepSeek also detailed its in-house search engine, which now powers both web/app search and a developer-facing Search & Fetch API.
This summary was generated automatically by AI from DeepSeek API Updates's publication. It is our own text, not a copy of the original — facts, figures and quotes belong to the source, linked above and below.
What changed?
- 1Flash model updated to DeepSeek-V4.1-Flash-1017, same API call method (deepseek-flash) and pricing
- 2Benchmark improvements across HLE, Terminal-Bench 4.0, DeepSWE v1.1, ProgramBench, NL2Repo-Bench, Automation-Bench, Agents' Last Exam, Apex Bench, Chartography, BabyVision, and ZeroBench-main
- 3Reduced task runtime and output length for lower cost and higher speed
- 4Specific optimizations for DeepSeek Harness, Tencent CodeBuddy, and WorkBuddy platforms; recommends DeepSeek Harness 0.2.1-rc.1 or later
- 5In-house DeepSeek search engine described: indexes hundreds of billions of pages, multilingual, filters SEO spam and low-quality AIGC
| Parameter | Before | Now |
|---|---|---|
| Model version | DeepSeek V4.1 | DeepSeek-V4.1-Flash-1017 |
| HLE (w/o tools) | Not specified | 42.6 (44.0 on pure-text subset) |
| Terminal-Bench 4.0 | Not specified | 42.3 |
| DeepSWE v1.1 | Not specified | 77.0 |
| Input/Output price | Unchanged | Unchanged |
| API access | deepseek-flash | deepseek-flash (same endpoint, updated model) |
Why it matters
For developers using DeepSeek's API, this is a free performance and efficiency upgrade with no pricing or interface changes, making it a low-friction improvement for agentic and coding workloads. The search API opening could also enable more reliable real-time information retrieval for AI agents.
What it means for AI agents and contact centers
If you use DeepSeek models in agent pipelines with tool calling or function execution, this update could lower latency and cost for the same price point — worth benchmarking against your current model version. The Search & Fetch API may also be relevant if your voice agents or automation flows need real-time web grounding.
🧪 Worth Testing
The model offers improved agentic and tool-use benchmarks at the same price with lower latency, and the new search API could add real-time retrieval capability to agent workflows — both are worth validating against current setups.
Sources
- DeepSeek API UpdatesOfficialPrimary sourceOriginal article →„Opening up the in-house search API“17 Oct 2026, 03:00
- Published by source
- 17 Oct 2026, 03:00
- Found by our system
- 17 Oct 2026, 11:12
- Summary generated
- 17 Oct 2026, 11:14
This article was written by AI from the original source. Facts, numbers and prices come from the source; missing values are marked “Not specified”. Legal notice, copyright and privacy