What is DeepSeek doing across hiring, forks, releases, and public writing?

Frontier labfresh 1d

DeepSeek

www.deepseek.com ↗github/deepseek-ai ↗dossier.json rss

Signal timeline110 total

Jun 28, 2026

Jun 28Modeldeepseek-ai/eagle3_gemma4_12b_ttt7New model variant release by DeepSeek.sourcenotability 5.0/10↓453♥12

Jun 28Modeldeepseek-ai/eagle3_qwen3_14b_ttt7Solid model release, no major traction evidence.sourcenotability 6.0/10↓378♥4

Jun 28Modeldeepseek-ai/eagle3_qwen3_8b_ttt7Solid DeepSeek fine-tune of Qwen3-8B with TTT, not a major launch.sourcenotability 5.0/10↓572♥3

Jun 28Modeldeepseek-ai/eagle3_qwen3_4b_ttt7Small DeepSeek model release, likely not a major launch.sourcenotability 5.0/10↓686♥4

Jun 28Modeldeepseek-ai/dflash_gemma4_12b_block7Minor model variant release, niche focus.sourcenotability 3.0/10↓273♥8

Jun 28Modeldeepseek-ai/dflash_qwen3_14b_block7New model release, not flagship.sourcenotability 5.0/10↓383♥6

Jun 28Modeldeepseek-ai/dflash_qwen3_8b_block7Notable model release from major labsourcenotability 7.0/10↓867♥4

Jun 28Modeldeepseek-ai/dflash_qwen3_4b_block7Minor release of a 4B parameter model variant.sourcenotability 5.0/10↓2k♥2

Jun 28Modeldeepseek-ai/dspark_gemma4_12b_block7Fine-tuned Gemma variant by top lab, moderate impact.sourcenotability 6.0/10↓73k♥45

Jun 28Modeldeepseek-ai/dspark_qwen3_14b_block7Specialized model variant release, limited public traction.sourcenotability 4.0/10↓8.6k♥14

Jun 28Modeldeepseek-ai/dspark_qwen3_8b_block7DeepSeek releases a niche 8B model variant, moderate expected impact.sourcenotability 5.0/10↓16k♥14

Jun 28Modeldeepseek-ai/dspark_qwen3_4b_block7DeepSeek's sparse Qwen3 variant release, notable research.sourcenotability 7.0/10↓9.9k♥50

Apr 22, 2026

Apr 22Modeldeepseek-ai/DeepSeek-V4-Pro-BaseNotable Pro base model release, moderate traction.sourcenotability 7.0/10↓28k♥336

Apr 22Modeldeepseek-ai/DeepSeek-V4-Flash-BaseNotable model release with strong traction; 67k downloadssourcenotability 8.0/10↓551k♥314

Jan 27, 2026

Jan 27Modeldeepseek-ai/DeepSeek-OCR-2High download count suggests strong community interest.sourcenotability 8.0/10↓940k♥1.1k

Dec 1, 2025

Dec 1Modeldeepseek-ai/DeepSeek-V3.2Major model release with exceptional community traction.sourcenotability 9.0/10↓1732k♥1.5k

Nov 28, 2025

Nov 28Modeldeepseek-ai/DeepSeek-V3.2-SpecialeNotable model release by DeepSeek with decent traction.sourcenotability 8.0/10↓5k♥724

Oct 17, 2025

Oct 17Modeldeepseek-ai/DeepSeek-OCR2.7M HF downloads, significant tractionsourcenotability 9.0/10↓2400k♥3.4k

Sep 29, 2025

Sep 29Modeldeepseek-ai/DeepSeek-V3.2-ExpMajor model release with high traction.sourcenotability 9.0/10↓60k♥1k

Sep 29Modeldeepseek-ai/DeepSeek-V3.2-Exp-BaseLow downloads, experimental model.sourcenotability 3.0/10↓617♥68

Aug 19, 2025

Aug 19Modeldeepseek-ai/DeepSeek-V3.1-BaseNotable model update from DeepSeeksourcenotability 7.0/10↓31k♥1k

May 29, 2025

May 29Modeldeepseek-ai/DeepSeek-R1-0528-Qwen3-8BVery high HF downloads for a notable fine-tune releasesourcenotability 8.0/10↓966k♥1.1k

May 28, 2025

May 28Modeldeepseek-ai/DeepSeek-R1-0528Frontier model release with massive tractionsourcenotability 10.0/10↓239k♥2.5k

Apr 30, 2025

Apr 30Modeldeepseek-ai/DeepSeek-Prover-V2-7BNotable model release from DeepSeek, moderate traction.sourcenotability 7.0/10↓97k♥147

Apr 30Modeldeepseek-ai/DeepSeek-Prover-V2-671BLarge model from DeepSeek but low traction.sourcenotability 7.0/10↓763♥831

Top signals

Agent answer

DeepSeek has 110 loaded public signals: 0 hiring, 0 forks, 71 releases or model cards, 0 talking, and 39 repos. Latest signal: deepseek-ai/deepseek-harness dsh-v0.1.5-rc.2. Data-business radar maps 7 signals to Data demand, Evals and quality, Infrastructure, Product and customer. The standing analysis was generated with deepseek-v4-pro and 94 evidence refs.

Hiring 0 Forks 0 Releases 71 Talking 0 Repos 39

DeepSeek

has loaded 110 public signals

DeepSeek

has hiring signal count 0

DeepSeek

has fork signal count 0

DeepSeek

has release signal count 71

Analysis — agent synthesisfull report →generated August 31, 2026

Thesis

DeepSeek's center of gravity is shifting from flagship frontier models toward agentic productization and efficiency infrastructure. The dominant recent artifact is DeepSeek Harness ('dsh'), an MIT-licensed TypeScript agent harness built on an 'everything is a plugin' Cordis architecture, which shipped seven tagged prereleases in roughly two weeks (v0.1.0-rc.7 → v0.1.2-alpha.2) P1 P2 P3 P4 P5 P6 P7 W3. Hiring and public framing confirm the pivot: 17 open roles oriented toward 'agent productization,' including a container scheduling/isolation platform and a Model Strategy PM Agent track, plus a stated goal to benchmark Claude Code and build a 'DeepSeek Code Harness' W1 W5. DeepSeek is simultaneously expanding its efficiency stack — DeepSpec for speculative decoding P21 E58, dspark/eagle3/dflash draft models E54 E55, and open infra (DeepEP, FlashMLA, DeepGEMM, 3FS, smallpond, DualPipe) E9 E11 E12 E14 E16 E28 — reinforced by the V4-Flash-0731 fine-tune beating the larger V4-Pro on independent tests W2 W4.

Signal desks

Hiring

Two Harness roles (Agent Harness product manager + research engineer, both Beijing-based) posted May 2026; researcher Chen Deli framed the goal as benchmarking Claude Code and building a 'DeepSeek Code Harness' W1.
17 recruitment positions; a full-stack developer role targets a 'next-generation container scheduling and isolation platform to support the operation of massive AI Agents' W5.
Model Strategy PM has a dedicated Agent track requiring familiarity with Claude Code, OpenClaw, and Manus, plus insight into personal assistants, Deep Research, automation workflows, and multi-modal device control W5.
Aug 1, 2026 internal-testing recruitment for open-source agent-harness developers drew 769 applicants, 712 unique repositories, and >1.2M cumulative GitHub stars W1.
No first-party job-post pages are in this pack; hiring is evidenced via secondary reporting W1 W5.

Forks

No cited evidence in this pack. Repos in the pack are marked 'Fork: no' P7 P21 P22 P24 P25 P26 P27 P28, and no fork events appear. The nearest adjacency is upstream integration — the Cordis plugin runtime P7 and Claude Code/Codex subagents P2 P5 P6 — not fork activity.

Releases

deepseek-harness prerelease train, v0.1.0-rc.7 (Aug 17) through v0.1.2-alpha.2 (Aug 30) P1 P2 P3 P4 P5 P6 E59 E60; highlights include the multimodal DeepSeek-V4-Flash-Vision-Exp adapter P4, Claude Code/Codex subagents P2 P5 P6, ACP/MCP/session control P2, Files API image uploads P3, a SQLite backend and 'DeepSeek Harness' trademark branding P5, and 'low' reasoning effort P6.
Model refreshes: DeepSeek-V4-Flash-0731 fine-tune (unchanged architecture, beats V4-Pro on independent tests) W2 W4; V4-Flash-Base (292B) and V4-Pro-Base (1.6T) released Apr 2026 E40 E41.
Multimodal/OCR: DeepSeek-OCR (2.44M downloads) E7; DeepSeek-OCR-2 (Apache-2.0, 1.1M downloads) E22 E57.
Efficiency/decoding: DeepSpec P21 E58; dspark/eagle3/dflash draft models over Qwen3 and Gemma4 bases E54 E55; DeepGEMM tag nv_dev_f8e8fb5 P8.
Frontier anchors: DeepSeek-R1 (7.29M downloads) E1; V3 / V3-0324 / R1-0528 / V3.2 E4 E8 E10 E20.

Talking

'Everything is a plugin': DeepSeek frames the harness as Agent = Model + Harness, with the model adapter, tool registry, session log, and agent loop all swappable plugins behind Cordis boundaries W3 W6.
Cost-efficiency: The Batch frames V4-Flash-0731 as surpassing the larger V4-Pro at a fraction of the cost of comparable proprietary models, small enough to run on the user's own hardware (3-bit) W2 W4.
Agent productization / Claude Code benchmarking narrative W1 W5.
HN attention: R1 1,843 points E2; DeepSeek-OCR 1,003 points E3; open-infra-index 757 E5; DeepEP 536 E9; FlashMLA 441 E11; DeepGEMM 391 E12.

Shipping

Two parallel shipping tracks dominate the window. (1) An agent runtime: deepseek-harness went from repo creation Aug 13 to seven tagged prereleases by Aug 30, adding Claude Code/Codex subagents, ACP/MCP session control, multimodal image handling, Files API uploads, persistent shells, and a SQLite backend P1 P2 P3 P4 P5 P6 P7. (2) Efficiency and model tooling: DeepSpec full-stack speculative-decoding training/eval P21 E58, draft models dspark/eagle3/dflash E54 E55, a DeepGEMM tag P8, and the V4-Flash-0731 fine-tune W2 W4. Historic anchors include R1, V3, OCR, and the R1-distill family E1 E4 E7 E17 E18 E29 E30 E32 E35.

Research themes

Speculative decoding / inference efficiency: DeepSpec (data prep, draft training, evaluation) P21 E58; dspark/eagle3/dflash draft models across Qwen3 and Gemma4 sizes E54 E55 P9 P10 P11 P12 P13 P14 P15 P16 P17 P18 P19 P20.
MoE, kernels, parallelism: DeepEP expert-parallel communication E9, FlashMLA multi-head latent attention kernels E11, DeepGEMM GPU BLAS kernels E12 P8, 3FS distributed filesystem E14, DualPipe bidirectional pipeline parallelism E28, LPLB expert-parallel load balancer E45, DeepSeek-MoE P22.
Multimodal/OCR/vision: DeepSeek-OCR/OCR-2 E3 E7 E22 E57, DeepSeek-VL/VL2 E39 E42 E46 E47, Janus/Janus-Pro E6 E38 E43, DreamCraft3D P24.
Reasoning/math/proving: R1/R1-Zero E1 E27, DeepSeek-Prover-V2 E13 E31 E48, DeepSeek-Math/Math-V2 E36 E56.
Code: DeepSeek-Coder/Coder-V2 and awesome-deepseek-coder P25 P23 E49 E51.

Hiring & scaling

Hiring is the clearest leading indicator of the agent pivot. Two Beijing Harness roles (PM + research engineer) and 17 open positions signal productization W1 W5. The full-stack role to build a 'container scheduling and isolation platform to support massive AI Agents' implies scaling of hosted agent workloads and sandbox infrastructure W5. The Model Strategy PM Agent track maps to GTM/product buildout across personal assistants, Deep Research, automation, and multi-modal device control W5. Beijing is the only named location in this pack W1.

Data-business implications

Data pipelines: smallpond (DuckDB + 3FS) and profile-data (computation-communication overlap analysis) are data-demand signals E16 E37; DeepSpec data prep warns of a ~38 TB target cache for the default Qwen3-4B setting, indicating heavy storage/data-engineering demand P21.
Infrastructure/deployment: 3FS (distributed FS for training/inference), DeepEP, FlashMLA, DeepGEMM, and DualPipe form a kernel/comms/storage layer E14 E9 E11 E12 E28; the harness adds a SQLite backend P5, and the roadmap points to a container scheduling/isolation platform W5.
Evals: DeepSpec is explicitly a training + evaluation codebase measuring speculative-decoding acceptance P21, flagged in the data radar as 'Evals and quality' E58.
Tooling/agent platform: deepseek-harness plugin architecture (Cordis), MCP/ACP, PTC mode, and Claude Code/Codex subagents create an agent-tooling integration surface P7 P2 P6 W3.
Safety: the Bubblewrap sandbox-escape fix and SAFETY.md update signal sandbox security hardening P4 P2.
Product/GTM: DeepSeek-OCR/OCR-2 map to document/data extraction surfaces E7 E22; the Model Strategy PM Agent track maps to personal-assistant, Deep Research, and automation GTM W5. No vendor or revenue claims are supported in this pack.

Traction highlights

deepseek-harness: 75,478 stars / 6,490 forks (Aug 14 metadata) P7; 205,346 stars logged at the creation event E24; >30,000 stars within hours of the leak W1.
DeepSeek-R1: 7,292,805 HF downloads, 13,417 likes E1; HN 1,843 points / 663 comments E2.
DeepSeek-OCR: 2,440,096 downloads E7; HN 1,003 points / 244 comments E3.
DeepSeek-V3: 104,440 GitHub stars E21; 1,067,717 model downloads E4.
Infra: FlashMLA 12,888 stars E11; 3FS 10,173 E14; DeepEP 10,080 E9; open-infra-index 8,060 E5; DeepGEMM 7,760 E12; smallpond 5,004 E16; DualPipe 2,998 E28.
Ecosystem/integration: awesome-deepseek-integration 39,003 stars E33; DeepSeek-Coder 23,659 stars P25.