DeepSeek 发布 V4.1-Flash-1017，DeepSWE 得分 77.0 并开放搜索 API
DeepSeek Just Released V4.1-Flash-1017, Scores 77.0 on DeepSWE, Opens Search API - MarkTechPost
DeepSeek 发布 DeepSeek-V4.1-Flash-1017，这是 V4.1-Flash 的新检查点，已在官网、App 和 API 上线，模型名为 deepseek-flash。DeepSeek 称其在自报的 11 项基准上均超过 9 月版 V4.1-Flash，提升 2.0 至 11.1 分，其中 DeepSWE v1.1 得 77.0，高于表中 Claude Opus 5 的 74.0，但 HLE 无工具为 42.6，低于 Claude Opus 5 的 56.3。同时开放的 Search 与 Fetch API 端点位于 api.deepseek.com，每个端点每 1000 次成功请求收费 0.75 美元，通过 DeepSeek Harness、Claude Code 或 Codex 发起的服务端调用在 2026 年 11 月 1 日前免费。
推荐理由
AI 评分 74
DeepSeek 新检查点在多项基准上超过前代，并同时开放搜索与抓取 API，读者可据此比较其与 Claude Opus 5 的强弱项。
主题
正文
当前显示原文，中文全文译文尚未就绪。
Add as a preferredsource on Google
DeepSeek has just released DeepSeek-V4.1-Flash-1017, a new checkpoint of its V4.1-Flash model. It is live on the DeepSeek website, app and API. DeepSeek claims stronger scores, faster task completion and shorter outputs than the September V4.1-Flash release. The company calls it a Pareto improvement in performance and cost. Alongside it, DeepSeek opened its Search and Fetch APIs to every developer. For agent builders, that means a cheaper coding model plus a first-party web layer behind one API key.
TL;aDR
Size: 552B backbone parameters plus 196B Engram parameters. It activates 8B parameters per token in prefill and 16B in decode. Context is 1M tokens, with up to 384K output tokens.
Runs on: DeepSeek API as deepseek-flash. September V4.1-Flash weights (763B parameters, FP8) are MIT-licensed on Hugging Face. Hardware requirements and 1017 weights: Not disclosed.
Performance: Beats the previous V4.1-Flash on all 11 benchmarks DeepSeek reported, by 2.0 to 11.1 points.
Best: 77.0 on DeepSWE v1.1, above Claude Opus 5 (74.0) in DeepSeek’s table.
Worst: 42.6 on HLE without tools, versus 56.3 for Claude Opus 5.
What is DeepSeek-V4.1-Flash-1017?
DeepSeek-V4.1-Flash-1017 is an updated post-training checkpoint of DeepSeek-V4.1-Flash, served through the deepseek-flash model name. The architecture details below come from the V4.1-Flash technical report.
DeepSeek team states that DeepSeek-V4.1-Flash-1017 uses reinforcement learning with self-feedback for end-to-end project work. Its examples run from game creation to 3D modeling. It is tuned for DeepSeek Harness, Tencent CodeBuddy and WorkBuddy. DeepSeek recommends DeepSeek Harness v0.2.1-rc.1 or later.
How does the V4.1-Flash architecture work?
V4.1-Flash is a multimodal Mixture-of-Experts model with 40 Transformer layers. They split into a 20-layer causal encoder and a 20-layer decoder.
Causal Encoder-Decoder (CED): Decoder global KV is projected from the final encoder hidden states. This nearly halves prefill compute.
Compressed Sparse Attention 2 (CSA2): Layers run in Full, Reindex or Reuse mode. They share main KV and indexer keys across layers.
FP4 main KV cache: Global KV drops to 890 bytes per token, about 1/4 of V4-Flash.
SWA Bounded Replay: Persistent KV cache shrinks to roughly 1/8 of V4-Flash.
Reasoning effort: A 1 to 100 scalar. The API tiers low, high and max map to 50, 75 and 100.
DeepSeek-V4.1-Flash-1017 benchmarks
All scores below are DeepSeek’s own figures. DeepSeek ran Terminal-Bench 4.0 in Claude Code and DeepSWE in Mini-SWE-V2. ZeroBench-main uses Pass@5.
BenchmarkV4.1-Flash-1017V4.1-FlashKimi K3GLM 5.3GPT 5.6-SolClaude Opus 5
DeepSWE v1.177.074.267.566.973.074.0
Terminal-Bench 4.042.331.212.637.939.951.8
Automation-Bench v1.0.660.554.846.748.845.850.3
Agents’ Last Exam35.431.827.628.526.728.6
Apex Bench v1.172.164.647.856.654.365.8
ProgramBench29.120.317.519.023.037.0
NL2Repo-Bench68.265.458.058.056.875.3
HLE (no tools)42.6 / 44.0*36.8 / 39.1*43.542.0*44.556.3
BabyVision (w/ tools)92.589.685.7Not disclosed88.994.1
*Text-only HLE subset.
The 1017 checkpoint leads on DeepSWE, Automation-Bench, Agents’ Last Exam and Apex Bench. Claude Opus 5 still leads on Terminal-Bench 4.0, ProgramBench, NL2Repo and HLE.
How does it compare with open-weight rivals?
FeatureDeepSeek-V4.1-Flash-1017Kimi K3GLM-5.3
Total parameters552B backbone + 196B Engram2.8T753B
Active parameters8B prefill / 16B decode104BNot disclosed
Context1M1,048,576Not disclosed
LicenseMIT (V4.1-Flash); 1017 weights Not disclosedKimi K3 LicenseGLM-5.3 license
ModalityText, image in; text outText, imageText
DeepSWE v1.177.067.566.9
Agents’ Last Exam35.428.328.5
API price (1M input miss / output, peak)$0.30 / $1.20Not disclosedNot disclosed
What are DeepSeek’s Search and Fetch APIs?
The Search API returns ranked web results for a query. The Fetch API returns a single page as Markdown. DeepSeek says its in-house engine serves hundreds of millions of requests daily.
Endpoints: POST /v1/tools/search and POST /v1/tools/fetch on api.deepseek.com.
Search options: 1 to 50 results, optional page content, and operators like site:, lang:, after: and before:.
Result fields: Each result has a 0 to 3 rerank score and a 0 to 4 authority tier.
Price: $0.75 per 1,000 successful requests for each endpoint. Empty results and failures are free.
Rate limit: 5 QPS per account for each API.
Free window: Server-side calls via DeepSeek Harness, Claude Code or Codex stay free until November 1, 2026.
Deployment and pricing
Call the model with deepseek-flash through OpenAI or Anthropic-format endpoints. Peak pricing is $0.30 per 1M cache-miss input tokens and $1.20 per 1M output tokens. Off-peak rates are half.
Key Takeaways
1017 scores 77.0 on DeepSWE v1.1, up from 74.2.
It leads Automation-Bench (60.5) and Agents’ Last Exam (35.4).
Claude Opus 5 still wins on HLE and Terminal-Bench 4.0.
Search and Fetch cost $0.75 per 1,000 calls each.
Agent harness calls stay free until November 1, 2026.
Asif RazzaqWebsite | + postsBio ⮌
RELATED ARTICLESMORE FROM AUTHOR
NVIDIA Releases Nemotron 3.5 Super VL, a 120B Open Multimodal Agent Model