https://derekwang2002.github.io/ai-intelligence/en/projects/deepseek/ Material changes in AI models, agents, engineering and infrastructure, with primary sources and recommendations. en Mon, 02 Nov 2026 16:56:00 GMT https://derekwang2002.github.io/ai-intelligence/en/events/ev-20260910-01/ ev-20260910-01-nvfp4 Sat, 31 Oct 2026 15:56:40 GMT Update TRIAL <p>NVIDIA publishes a 1017 NVFP4 checkpoint for Blackwell GPUs</p><p>Original story: DeepSeek V4.1-Flash: new CED/CSA2 architecture with 1M context, open weights</p><p><strong>Why read: </strong>NVIDIA’s NVFP4 checkpoint lowers the Blackwell deployment barrier, but quantized quality still needs testing.</p><p>Recommendation: TRIAL</p><p>Sources: <a href="https://build.nvidia.com/nvidia/deepseek-v4_1-flash-1017/modelcard">DeepSeek V4.1 Flash 1017 NVFP4 model card</a></p> https://derekwang2002.github.io/ai-intelligence/en/events/ev-20260910-01/ ev-20260910-01-aa-1017 Mon, 19 Oct 2026 16:04:44 GMT Update TRIAL <p>Artificial Analysis independently measures 1017 speed, task cost and verbosity</p><p>Original story: DeepSeek V4.1-Flash: new CED/CSA2 architecture with 1M context, open weights</p><p>Recommendation: TRIAL</p><p>Sources: <a href="https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017">DeepSeek V4.1 Flash 1017 Max independent analysis</a></p> https://derekwang2002.github.io/ai-intelligence/en/events/ev-20261019-03/ ev-20261019-03-published Mon, 19 Oct 2026 16:04:44 GMT New WATCH <p>The U.S. Department of Homeland Security publicly filed notice of an exclusion and removal order covering DeepSeek, affiliated entities and third-party products or services developed using DeepSeek implementations in federal civilian agency IT. The signed order is dated July 24, 2026 and required exclusion and removal immediately upon issuance; the Federal Register notice is scheduled for October 20. Agencies may seek written exceptions within 60 days.</p><p><strong>Why read: </strong>The order reaches third-party DeepSeek implementations, requiring a federal supply-chain inventory without generalizing it into a universal ban.</p><p>Recommendation: WATCH</p><p>Sources: <a href="https://public-inspection.federalregister.gov/2026-21447.pdf">DHS DeepSeek exclusion and removal order notice</a></p> https://derekwang2002.github.io/ai-intelligence/en/events/ev-20260910-01/ ev-20260910-01-1017 Sat, 17 Oct 2026 15:59:50 GMT Update TRIAL <p>V4.1-Flash-1017 weights and hosted update go live</p><p>Original story: DeepSeek V4.1-Flash: new CED/CSA2 architecture with 1M context, open weights</p><p><strong>Why read: </strong>Weights and the hosted alias update without an API migration, but vendor benchmarks need same-task reproduction.</p><p>Recommendation: TRIAL</p><p>Sources: <a href="https://api-docs.deepseek.com/news/news261017">DeepSeek V4.1-Flash-1017 release</a> · <a href="https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017">DeepSeek-V4.1-Flash-1017 weights</a></p> https://derekwang2002.github.io/ai-intelligence/en/events/ev-20260910-01/ ev-20260910-01-published Fri, 11 Sep 2026 18:31:49 GMT New TRIAL <p>DeepSeek V4.1-Flash is an open-weight multimodal model with 1M context and a sparse CED/CSA2 architecture. The October 17 1017 update replaced the website, app and deepseek-flash API alias without changing price or interface and shipped a new MIT checkpoint. Artificial Analysis now independently scores the hosted Max configuration at 45 on its Intelligence Index, $0.26 per task, 224.7 output tokens per second and 1.12-second TTFT; the run used 210M output tokens, so high verbosity remains a cost caveat.</p><p>Recommendation: TRIAL</p><p>Sources: <a href="https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash">huggingface.co · DeepSeek V4.1 Flash</a></p> https://derekwang2002.github.io/ai-intelligence/en/events/ev-20260831-02/ ev-20260831-02-published Wed, 02 Sep 2026 00:01:14 GMT New WATCH <p>DeepSeek released DeepSeek-V4-Flash-Vision-Exp weights on Hugging Face under MIT: its first experimental multimodal model in the V4 family, built on the DeepSeek-V4-Flash architecture with added visual modules and continued training. The model card benchmarks it against V4-Flash-0731 and Opus-4.8: Terminal Bench 2.1 83.9 (V4-Flash 82.7, Opus-4.8 85.0), DeepSWE 59.3 (54.4 / 58.0), and a jump on multimodal agent benchmarks — ApexBench 36.5 vs 26.2 for the text-only sibling (which ignores multimodal input). The API had been live on the DeepSeek platform since ~8/21 (set model to deepseek-v4-flash-vision-exp); the 8/31 release adds the weights and a minimal PyTorch reference implementation.</p><p>Recommendation: WATCH</p><p>Sources: <a href="https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-Vision-Exp">huggingface.co · DeepSeek V4 Flash Vision Exp</a></p> https://derekwang2002.github.io/ai-intelligence/en/events/ev-20260813-02/ ev-20260813-02-published Sat, 15 Aug 2026 12:26:05 GMT New TRIAL <p>DeepSeek moved V4-Pro from preview to GA across app, web, and API (deepseek-v4-pro), with native OpenAI Responses API support and a one-click Codex configuration. Thinking-effort levels (low/high/max) were added for V4-Pro and V4-Flash. Peak/off-peak pricing takes effect 2026-08-16T16:00Z, with off-peak at 50% of peak.</p><p>Recommendation: TRIAL</p><p>Sources: <a href="https://api-docs.deepseek.com/updates">api-docs.deepseek.com · updates</a></p> https://derekwang2002.github.io/ai-intelligence/en/events/ev-20260814-05/ ev-20260814-05-published Sat, 15 Aug 2026 12:26:05 GMT New WATCH <p>DeepSeek open-sourced its agent harness — the framework behind its own benchmark runs — under MIT. It is a TypeScript/Node monorepo built on the Cordis framework with an everything-is-a-plugin architecture, positioned as an open challenge to Claude Code. It runs via npx @deepseek-ai/dsh web, with the Web UI at 127.0.0.1:3080. Config is compatible with AGENTS.md/CLAUDE.md. This is a developer preview, and DeepSeek explicitly warns of breaking changes.</p><p>Recommendation: WATCH</p><p>Sources: <a href="https://github.com/deepseek-ai/deepseek-harness">deepseek-ai/deepseek-harness · GitHub</a></p>