DeepSeek(@deepseek_ai)

🌐 Supporting open source. Expanding deployment options. We’ll work closely with the open-source co...

DeepSeek (@deepseek_ai) — Thu, 10 Sep 2026 06:10:14 GMT

🌐 Supporting open source. Expanding deployment options.

We’ll work closely with the open-source community on V4.1-Flash inference support and explore more deployment options.
Planning a large-scale deployment with 2,000 GPUs + a storage cluster? Let’s talk.

🔹 Modelhuggingface.co/deepseek-ai/De…6a
🔹 Papehuggingface.co/deepseek-ai/De…SSV

6/6

💬66🔄210❤️2787👀328290📊366

⚡ Powered by xgo.ing

💰 More efficient architecture. Lower API prices. V4.1-Flash lets us serve more users at a lower co...

DeepSeek (@deepseek_ai) — Thu, 10 Sep 2026 06:10:13 GMT

💰 More efficient architecture. Lower API prices.

V4.1-Flash lets us serve more users at a lower cost. We’re passing the savings on to you.

🔹 Peak/off-peak pricing continues to balance demand.
🔹 Off-peak rates are 50% of peak rates. Schedule flexible workloads off-peak to save.
🔹 New pricing takes effect at 04:00 UTC on Sept 10, 2026.

5/6

💬21🔄51❤️1811👀226887📊181

⚡ Powered by xgo.ing

⚡ V4.1-Flash is now live on the DeepSeek API with native multimodal support. Set your model to deep...

DeepSeek (@deepseek_ai) — Thu, 10 Sep 2026 06:10:12 GMT

⚡ V4.1-Flash is now live on the DeepSeek API with native multimodal support.

Set your model to deepseek-flash.

🔹 V4-Flash & V4-Flash-Vision-Exp are retired. For compatibility, deepseek-v4-flash and deepseek-v4-flash-vision-exp temporarily route to V4.1-Flash.
🔹 Tests by multiple parties put V4.1-Flash ahead of V4-Pro on performance, cost, speed & total runtime. We’re phasing out V4-Pro.
🔹 Starting at 04:00 UTC on Sept 14, 2026, all deepseek-v4-pro requests will route to V4.1-Flash at V4.1-Flash rates. This will continue until V4.1-Pro launches.

🤝 Official partners @WorkBuddy_AI  (including Codebuddy) & @opencode  now fully support V4.1-Flash. Try it today!

4/6

💬20🔄52❤️1936👀387408📊187

⚡ Powered by xgo.ing

💾 Smaller KV cache. Bigger savings. Compared with the previous generation, V4.1-Flash’s KV cache n...

DeepSeek (@deepseek_ai) — Thu, 10 Sep 2026 06:10:11 GMT

💾 Smaller KV cache. Bigger savings.

Compared with the previous generation, V4.1-Flash’s KV cache needs just:
🔹 1/4 the HBM
🔹 1/8 the SSD storage
Cache-hit charges often account for a large share of agent costs. Compressing the cache cuts those costs significantly.

3/6

💬29🔄186❤️2968👀761886📊354

⚡ Powered by xgo.ing

🧠 Asymmetric architecture. More intelligence, less cost. 🔹 552B-parameter MoE. 🔹 New Causal Enco...

DeepSeek (@deepseek_ai) — Thu, 10 Sep 2026 06:10:10 GMT

🧠 Asymmetric architecture. More intelligence, less cost.

🔹 552B-parameter MoE.
🔹 New Causal Encoder–Decoder architecture: just 8B active parameters for input, 16B for output.
🔹 New pre-training methods + larger-scale RL post-training deliver benchmark results ahead of flagship models, including DeepSeek-V4-Pro.

2/6

💬88🔄345❤️4627👀1590506📊688

⚡ Powered by xgo.ing

🚀 Introducing DeepSeek-V4.1-Flash: smarter, faster, more efficient. 🔹 Introducing the smallest mo...

DeepSeek (@deepseek_ai) — Thu, 10 Sep 2026 06:10:09 GMT

🚀 Introducing DeepSeek-V4.1-Flash: smarter, faster, more efficient.

🔹 Introducing the smallest model in our new architecture family, with native visual understanding.
🔹 Designed for greater capability, faster inference, higher throughput, and scaling to larger models.

1/6

💬912🔄2994❤️27376👀5804108📊4275

⚡ Powered by xgo.ing

Multimodal API support 🔌 🔹 Set model='deepseek-v4-flash-vision-exp' 🔹 Images are tokenized for b...

DeepSeek (@deepseek_ai) — Fri, 21 Aug 2026 09:17:40 GMT

Multimodal API support 🔌

🔹 Set model='deepseek-v4-flash-vision-exp'
🔹 Images are tokenized for billing: up to 384 tokens each, at V4-Flash pricing
🔹 Supports Chat Completions, Messages & Responses
🔹 Supports mixed text + image input; images can be provided via base64, external URLs, or the Files API.

Dapi-docs.deepseek.com/guides/visionZ3wWB

3/n

💬7🔄18❤️554👀110106📊57

⚡ Powered by xgo.ing

Files API is now live. 📁 🔹 Free to use 🔹 Upload an image once, then reference it by file_id to s...

DeepSeek (@deepseek_ai) — Fri, 21 Aug 2026 09:17:40 GMT

Files API is now live. 📁

🔹 Free to use
🔹 Upload an image once, then reference it by file_id to save request bandwidth
🔹 Reuse the same image across requests—no need to upload it again

Learn moapi-docs.deepseek.com/guides/files_a…6xmc

4/n

💬9🔄19❤️517👀97867📊59

⚡ Powered by xgo.ing

Multimodality unlocks more agent use cases. 👀 V4-Flash-Vision-Exp works smoothly across agent fram...

DeepSeek (@deepseek_ai) — Fri, 21 Aug 2026 09:17:39 GMT

Multimodality unlocks more agent use cases. 👀

V4-Flash-Vision-Exp works smoothly across agent frameworks, combining visual understanding with a wide range of tools to unlock more practical workflows.

2/n

💬12🔄19❤️629👀65950📊64

⚡ Powered by xgo.ing

DeepSeek-V4-Flash-Vision-Exp is now live on the DeepSeek API Platform! 🚀 🔹 This experimental mult...

DeepSeek (@deepseek_ai) — Fri, 21 Aug 2026 09:17:38 GMT

DeepSeek-V4-Flash-Vision-Exp is now live on the DeepSeek API Platform! 🚀

🔹 This experimental multimodal model matches DeepSeek-V4-Flash on text capabilities—including agents, reasoning, and world knowledge.
🔹 On multimodal agent benchmarks, V4-Flash-Vision-Exp makes a major leap over V4-Flash, bringing multimodal agent performance close to Opus-4.8.

Try it with model='deepseek-v4-flash-vision-exp'. DeepSeek Harness 0.1.1 was released today with out-of-the-box support for the new model.

1/n

💬559🔄1221❤️11500👀2697084📊1929

⚡ Powered by xgo.ing

🧩 DeepSeek Harness v0.1 is now available in Developer Preview! 🔹 We’re opening it up to developer...

DeepSeek (@deepseek_ai) — Thu, 13 Aug 2026 13:02:03 GMT

🧩 DeepSeek Harness v0.1 is now available in Developer Preview!

🔹 We’re opening it up to developers building agent harnesses worldwide and open-sourcing the codebase in MIT license.
🔹 Powered by the Cordis meta-framework, DeepSeek Harness is an agent harness built around one core idea: Everything is a plugin. Models, tools, skills, sessions, sandboxes, filesystems, loops, orchestration, and UI are ALL implemented as plugins, and can be mixed, matched, replaced, and extended.

Try it nogithub.com/deepseek-ai/de…hKA

💬842🔄2465❤️20124👀4431410📊3937

⚡ Powered by xgo.ing

API pricing update 💰 With the V4 lineup release, we’re updating our API pricing and introducing pe...

DeepSeek (@deepseek_ai) — Thu, 13 Aug 2026 11:31:23 GMT

API pricing update 💰

With the V4 lineup release, we’re updating our API pricing and introducing peak and off-peak rates. Off-peak rates are 50% lower than peak, enabling more flexible workload scheduling. 📉

New pricing takes effect at 16:00 UTC, Aug 16, 2026 🕒

💬203🔄192❤️1823👀585103📊430

⚡ Powered by xgo.ing

We’re launching DeepSeek-V4-Pro today! 🚀 🔷 Major Agent upgrades with strong production gains! 🔷 ...

DeepSeek (@deepseek_ai) — Thu, 13 Aug 2026 11:31:22 GMT

We’re launching DeepSeek-V4-Pro today! 🚀

🔷 Major Agent upgrades with strong production gains!
🔷 Flexible reasoning effort for V4-Pro & V4-Flash: low for simple tasks, high for daily Agent workflows, max for complex tasks.
🔷 Native OpenAI Responses API support, optimized for Codex with one-click setup.

V4 Pro is now available on app/web. Try it via “Expert Mode”.
V4 Pro is also available via API. Model names remain unchanged—please refer to the API docs for setup details.

💬615🔄1802❤️15341👀3025372📊2299

⚡ Powered by xgo.ing

⚠️ Note 🔷 DeepSeek-V4-Flash-0731 keeps the exact same model architecture and size as the preview v...

DeepSeek (@deepseek_ai) — Fri, 31 Jul 2026 06:56:42 GMT

⚠️ Note

🔷 DeepSeek-V4-Flash-0731 keeps the exact same model architecture and size as the preview version.
🔷 Today's upgrade applies ONLY to the DeepSeek-V4-Flash API. The DeepSeek-V4-Pro API and App/Web models remain unchanged for now.

The official release of DeepSeek-V4-Pro is coming ASAP! Stay tuned.

💬92🔄139❤️3834👀549086📊405

⚡ Powered by xgo.ing

🚀 DeepSeek-V4-Flash Official API is now LIVE in public beta! 🔷 We’ve massively upgraded its Agent...

DeepSeek (@deepseek_ai) — Fri, 31 Jul 2026 06:56:41 GMT

🚀 DeepSeek-V4-Flash Official API is now LIVE in public beta!

🔷 We’ve massively upgraded its Agent capabilities—benchmark scores are now far surpassing the V4-Pro-Preview. Check out the massive performance leap below! 👇
🔷 The official V4-Flash now natively supports the Responses API format and is fully adapted for Codex!

Check out the configuration details in our official API doapi-docs.deepseek.com/quick_start/ag…Meiq

💬1716🔄3535❤️30016👀9562589📊5330

⚡ Powered by xgo.ing

We are making our discount permanent! 🎉 Enjoy building with DeepSeek-V4-Pro and bring your innovat...

DeepSeek (@deepseek_ai) — Fri, 22 May 2026 16:01:03 GMT

We are making our discount permanent! 🎉

Enjoy building with DeepSeek-V4-Pro and bring your innovative ideas to life! 🚀

DeepSeek@deepseek_ai

The DeepSeek-V4-Pro discount has been extended until May 31, 2026, 15:59 UTC!

🔗 View Quoted Tweet

💬1485🔄2748❤️24336👀7084632📊3635

⚡ Powered by xgo.ing

The DeepSeek-V4-Pro discount has been extended until May 31, 2026, 15:59 UTC!

DeepSeek (@deepseek_ai) — Wed, 29 Apr 2026 02:20:52 GMT

The DeepSeek-V4-Pro discount has been extended until May 31, 2026, 15:59 UTC!

DeepSeek@deepseek_ai

🔥DeepSeek-V4-Pro API is 75% OFF until May 5th, 2026, 15:59 (UTC Time)! Don't miss out on this massive discount.

🛠️Integration Updates:
🔹Claude Code: Set model to deepseek-v4-pro[1m]  to unlock 1M context!
🔹OpenCode: Update to v1.14.24+
🔹OpenClaw: Update to v2026.4.24+

Check the latest official API docs for full detaapi-docs.deepseek.com/quick_start/pr…dDpyU

🔗 View Quoted Tweet

💬321🔄556❤️6807👀2752030📊802

⚡ Powered by xgo.ing

🔥DeepSeek Input Cache Price Drop! Effective immediately, the price for input cache hits across the...

DeepSeek (@deepseek_ai) — Sun, 26 Apr 2026 16:35:11 GMT

🔥DeepSeek Input Cache Price Drop!

Effective immediately, the price for input cache hits across the ENTIRE DeepSeek API series is reduced to just 1/10th of the original price! Build more efficiently for less.

📌Reminder: The DeepSeek-V4-Pro 75% OFF promotion is still active until May 5th, 2026, 15:59 (UTC Time).

💬606🔄784❤️8862👀1686401📊1467

⚡ Powered by xgo.ing

🔥DeepSeek-V4-Pro API is 75% OFF until May 5th, 2026, 15:59 (UTC Time)! Don't miss out on this massi...

DeepSeek (@deepseek_ai) — Sat, 25 Apr 2026 15:33:11 GMT

🔥DeepSeek-V4-Pro API is 75% OFF until May 5th, 2026, 15:59 (UTC Time)! Don't miss out on this massive discount.

🛠️Integration Updates:
🔹Claude Code: Set model to deepseek-v4-pro[1m]  to unlock 1M context!
🔹OpenCode: Update to v1.14.24+
🔹OpenClaw: Update to v2026.4.24+

Check the latest official API docs for full detaapi-docs.deepseek.com/quick_start/pr…dDpyU

💬349🔄882❤️9277👀1436079📊1455

⚡ Powered by xgo.ing

🔹 Amid recent attention, a quick reminder: please rely only on our official accounts for DeepSeek n...

DeepSeek (@deepseek_ai) — Fri, 24 Apr 2026 03:24:16 GMT

🔹 Amid recent attention, a quick reminder: please rely only on our official accounts for DeepSeek news. Statements from other channels do not reflect our views.
🔹 Thank you for your continued trust. We remain committed to longtermism, advancing steadily toward our ultimate goal of AGI.

7/n

💬32🔄86❤️2180👀330963📊209

⚡ Powered by xgo.ing

API is Available Today! 🔹 Keep base_url, just update model to deepseek-v4-pro or deepseek-v4-flash...

DeepSeek (@deepseek_ai) — Fri, 24 Apr 2026 03:24:15 GMT

API is Available Today!

🔹 Keep base_url, just update model to deepseek-v4-pro or deepseek-v4-flash.
🔹 Supports OpenAI ChatCompletions & Anthropic APIs.
🔹 Both models support 1M context & dual modes (Thinking / Non-Thinkingapi-docs.deepseek.com/guides/thinkin…I8T

⚠️ Note: deepseek-chat & deepseek-reasoner will be fully retired and inaccessible after Jul 24th, 2026, 15:59 (UTC Time). (Currently routing to deepseek-v4-flash non-thinking/thinking).

6/n

💬31🔄126❤️2094👀435271📊236

⚡ Powered by xgo.ing

Structural Innovation & Ultra-High Context Efficiency 🔹 Novel Attention: Token-wise compression + ...

DeepSeek (@deepseek_ai) — Fri, 24 Apr 2026 03:24:13 GMT

Structural Innovation & Ultra-High Context Efficiency

🔹 Novel Attention: Token-wise compression + DSA (DeepSeek Sparse Attention).
🔹 Peak Efficiency: World-leading long context with drastically reduced compute & memory costs.
🔹 1M Standard: 1M context is now the default across all official DeepSeek services.

4/n

💬20🔄140❤️2080👀664400📊230

⚡ Powered by xgo.ing

DeepSeek-V4-Flash 🔹 Reasoning capabilities closely approach V4-Pro. 🔹 Performs on par with V4-Pro...

DeepSeek (@deepseek_ai) — Fri, 24 Apr 2026 03:24:12 GMT

DeepSeek-V4-Flash

🔹 Reasoning capabilities closely approach V4-Pro.
🔹 Performs on par with V4-Pro on simple Agent tasks.
🔹 Smaller parameter size, faster response times, and highly cost-effective API pricing.

3/n

💬17🔄137❤️2104👀411656📊220

⚡ Powered by xgo.ing

DeepSeek-V4-Pro 🔹 Enhanced Agentic Capabilities: Open-source SOTA in Agentic Coding benchmarks. ...

DeepSeek (@deepseek_ai) — Fri, 24 Apr 2026 03:24:10 GMT

DeepSeek-V4-Pro

🔹 Enhanced Agentic Capabilities: Open-source SOTA in Agentic Coding benchmarks.
🔹 Rich World Knowledge: Leads all current open models, trailing only Gemini-3.1-Pro.
🔹 World-Class Reasoning: Beats all current open models in Math/STEM/Coding, rivaling top closed-source models.

2/n

💬52🔄320❤️3588👀608628📊425

⚡ Powered by xgo.ing

🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M co...

DeepSeek (@deepseek_ai) — Fri, 24 Apr 2026 03:24:09 GMT

🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length.

🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models.
🔹 DeepSeek-V4-Flash: 284B total / 13B active params. Your fast, efficient, and economical choice.

Try it now chat.deepseek.com1Dl via Expert Mode / Instant Mode. API is updated & available today!

📄 Tech Repohuggingface.co/deepseek-ai/De…kYtp
🤗 Open Weighuggingface.co/collections/de…i7SDM

1/n

💬1665🔄7499❤️45644👀10101434📊7991

⚡ Powered by xgo.ing

🛠 Open Source Release 📦 DeepSeek-V3.2 Model: https://t.co/Kh8HzHl3uX 📦 DeepSeek-V3.2-Speciale Mo...

DeepSeek (@deepseek_ai) — Mon, 01 Dec 2025 11:19:22 GMT

🛠 Open Source Release

📦 DeepSeek-V3.2 Modelhuggingface.co/deepseek-ai/De…uX
📦 DeepSeek-V3.2-Speciale Modehuggingface.co/deepseek-ai/De…0AJ
📄 Tech repohuggingface.co/deepseek-ai/De…lMpX

5/n

💬48🔄133❤️1568👀308162📊213

⚡ Powered by xgo.ing

💻 API Update 🔹 V3.2: Same usage pattern as V3.2-Exp. 🔹 V3.2-Speciale: Served via a temporary end...

DeepSeek (@deepseek_ai) — Mon, 01 Dec 2025 11:19:21 GMT

💻 API Update

🔹 V3.2: Same usage pattern as V3.2-Exp.
🔹 V3.2-Speciale: Served via a temporary endpoint: base_urapi.deepseek.com/v3.2_speciale_…Lbm". Same pricing as V3.2, no tool calls, available until Dec 15th, 2025, 15:59 (UTC Time).
💡 V3.2 now supports Thinking in Tool-Use — detaiapi-docs.deepseek.com/guides/thinkin…DI8T

4/n

💬39🔄77❤️1252👀426005📊151

⚡ Powered by xgo.ing

🚀 Launching DeepSeek-V3.2 & DeepSeek-V3.2-Speciale — Reasoning-first models built for agents! 🔹 D...

DeepSeek (@deepseek_ai) — Mon, 01 Dec 2025 11:19:18 GMT

🚀 Launching DeepSeek-V3.2 & DeepSeek-V3.2-Speciale — Reasoning-first models built for agents!

🔹 DeepSeek-V3.2: Official successor to V3.2-Exp. Now live on App, Web & API.
🔹 DeepSeek-V3.2-Speciale: Pushing the boundaries of reasoning capabilities. API-only for now.

📄 Tech repohuggingface.co/deepseek-ai/De…NuG0

1/n

💬944🔄2403❤️16000👀5325924📊3103

⚡ Powered by xgo.ing

⚠️ Heads-up to anyone using the DeepSeek-V3.2-Exp inference demo: earlier versions had a RoPE implem...

DeepSeek (@deepseek_ai) — Tue, 18 Nov 2025 03:00:56 GMT

⚠️ Heads-up to anyone using the DeepSeek-V3.2-Exp inference demo: earlier versions had a RoPE implementation mismatch in the indexer module that could degrade performance. Indexer RoPE expects non-interleaved input, MLA RoPE expects interleaved. Fixed in github.com/deepseek-ai/De….

💬228🔄167❤️2252👀735848📊405

⚡ Powered by xgo.ing

🛠 Open Source Release 🔗 Model: https://t.co/kORJG3nCWN 🔗 Tech report: https://t.co/X8Wcqbhg5a ...

DeepSeek (@deepseek_ai) — Mon, 29 Sep 2025 10:10:03 GMT

🛠 Open Source Release

🔗 Modelhuggingface.co/deepseek-ai/De…WN
🔗 Tech reporgithub.com/deepseek-ai/De…g5a
🔗 Key GPU kernels in TileLang & CUDA (use TileLang for rapid research prototyping!)

4/n

💬26🔄103❤️876👀235870📊125

⚡ Powered by xgo.ing

💻 API Update 🎉 Lower costs, same access! 💰 DeepSeek API prices drop 50%+, effective immediately....

DeepSeek (@deepseek_ai) — Mon, 29 Sep 2025 10:10:02 GMT

💻 API Update

🎉 Lower costs, same access!
💰 DeepSeek API prices drop 50%+, effective immediately.

🔹 For comparison testing, V3.1-Terminus remains available via a temporary API until Oct 15th, 2025, 15:59 (UTC Time). Detaiapi-docs.deepseek.com/guides/compari…KylK
🔹 Feedback welcfeedback.deepseek.com/dsa7FDPr

3/n

💬33🔄92❤️1068👀393486📊143

⚡ Powered by xgo.ing

🚀 Introducing DeepSeek-V3.2-Exp — our latest experimental model! ✨ Built on V3.1-Terminus, it debu...

DeepSeek (@deepseek_ai) — Mon, 29 Sep 2025 10:10:00 GMT

🚀 Introducing DeepSeek-V3.2-Exp — our latest experimental model!

✨ Built on V3.1-Terminus, it debuts DeepSeek Sparse Attention(DSA) for faster, more efficient training & inference on long context.
👉 Now live on App, Web, and API.
💰 API prices cut by 50%+!

1/n

💬323🔄876❤️6951👀1418023📊1145

⚡ Powered by xgo.ing

🚀 DeepSeek-V3.1 → DeepSeek-V3.1-Terminus The latest update builds on V3.1’s strengths while address...

DeepSeek (@deepseek_ai) — Mon, 22 Sep 2025 13:27:43 GMT

🚀 DeepSeek-V3.1 → DeepSeek-V3.1-Terminus
The latest update builds on V3.1’s strengths while addressing key user feedback.

✨ What’s improved?
🌐 Language consistency: fewer CN/EN mix-ups & no more random chars.
🤖 Agent upgrades: stronger Code Agent & Search Agent performance.

1/2

💬182🔄518❤️4573👀840297📊686

⚡ Powered by xgo.ing

Introducing DeepSeek-V3.1: our first step toward the agent era! 🚀 🧠 Hybrid inference: Think & Non...

DeepSeek (@deepseek_ai) — Thu, 21 Aug 2025 06:33:07 GMT

Introducing DeepSeek-V3.1: our first step toward the agent era! 🚀

🧠 Hybrid inference: Think & Non-Think — one model, two modes
⚡️ Faster thinking: DeepSeek-V3.1-Think reaches answers in less time vs. DeepSeek-R1-0528
🛠️ Stronger agent skills: Post-training boosts tool use and multi-step agent tasks

Try it now — toggle Think/Non-Think via the "DeepThink" buttochat.deepseek.comEz2

1/5

💬504🔄1741❤️14701👀2095477📊2259

⚡ Powered by xgo.ing

🚀 DeepSeek-R1-0528 is here! 🔹 Improved benchmark performance 🔹 Enhanced front-end capabilities ...

DeepSeek (@deepseek_ai) — Thu, 29 May 2025 12:11:19 GMT

🚀 DeepSeek-R1-0528 is here!

🔹 Improved benchmark performance
🔹 Enhanced front-end capabilities
🔹 Reduced hallucinations
🔹 Supports JSON output & function calling

✅ Try it chat.deepseek.comcbEz2
🔌 No change to API usage — docs api-docs.deepseek.com/guides/reasoni…6et35u
🔗 Open-source wehuggingface.co/deepseek-ai/De…yhHptsC

💬552🔄1463❤️9626👀1491396📊1661

⚡ Powered by xgo.ing

🚀 DeepSeek-V3-0324 is out now! 🔹 Major boost in reasoning performance 🔹 Stronger front-end devel...

DeepSeek (@deepseek_ai) — Tue, 25 Mar 2025 13:32:43 GMT

🚀 DeepSeek-V3-0324 is out now!

🔹 Major boost in reasoning performance
🔹 Stronger front-end development skills
🔹 Smarter tool-use capabilities

✅ For non-complex reasoning tasks, we recommend using V3 — just turn off “DeepThink”
🔌 API usage remains unchanged
📜 Models are now released under the MIT License, just like DeepSeek-R1!
🔗 Open-source wehuggingface.co/deepseek-ai/De…MLhfngk

💬674🔄1869❤️11602👀1566323📊2003

⚡ Powered by xgo.ing

🚀 Day 6 of #OpenSourceWeek: One More Thing – DeepSeek-V3/R1 Inference System Overview Optimized th...

DeepSeek (@deepseek_ai) — Sat, 01 Mar 2025 04:11:25 GMT

🚀 Day 6 of#OpenSourceWeekk: One More Thing – DeepSeek-V3/R1 Inference System Overview

Optimized throughput and latency via:
🔧 Cross-node EP-powered batch scaling
🔄 Computation-communication overlap
⚖️ Load balancing

Statistics of DeepSeek's Online Service:
⚡ 73.7k/14.8k input/output tokens per second per H800 node
🚀 Cost profit margin 545%

💡 We hope this week's insights offer value to the community and contribute to our shared AGI goals.
📖 Deep bit.ly/4ihZUiO3mCxF5

💬774🔄1191❤️9038👀3981840📊1832

⚡ Powered by xgo.ing

🚀 Day 5 of #OpenSourceWeek: 3FS, Thruster for All DeepSeek Data Access Fire-Flyer File System (3FS...

DeepSeek (@deepseek_ai) — Fri, 28 Feb 2025 01:06:38 GMT

🚀 Day 5 of#OpenSourceWeekk: 3FS, Thruster for All DeepSeek Data Access

Fire-Flyer File System (3FS) - a parallel file system that utilizes the full bandwidth of modern SSDs and RDMA networks.

⚡ 6.6 TiB/s aggregate read throughput in a 180-node cluster
⚡ 3.66 TiB/min throughput on GraySort benchmark in a 25-node cluster
⚡ 40+ GiB/s peak throughput per client node for KVCache lookup
🧬 Disaggregated architecture with strong consistency semantics
✅ Training data preprocessing, dataset loading, checkpoint saving/reloading, embedding vector search & KVCache lookups for inference in V3/R1

📥 3FSgithub.com/deepseek-ai/3FS1aK
⛲ Smallpond - data processing framework on 3FSgithub.com/deepseek-ai/sm…jZK

💬522🔄1225❤️10146👀3214227📊1876

⚡ Powered by xgo.ing

🚀 Day 4 of #OpenSourceWeek: Optimized Parallelism Strategies ✅ DualPipe - a bidirectional pipeline...

DeepSeek (@deepseek_ai) — Thu, 27 Feb 2025 02:05:53 GMT

🚀 Day 4 of#OpenSourceWeekk: Optimized Parallelism Strategies

✅ DualPipe - a bidirectional pipeline parallelism algorithm for computation-communication overlap in V3/R1 training.
�github.com/deepseek-ai/Du…po

✅ EPLB - an expert-parallel load balancer for V3/R1.
github.com/deepseek-ai/ep…n49

📊 Analyze computation-communication overlap in V3/R1github.com/deepseek-ai/pr…jA1Lw

💬441🔄791❤️5793👀2551343📊1026

⚡ Powered by xgo.ing

🚀 Day 3 of #OpenSourceWeek: DeepGEMM Introducing DeepGEMM - an FP8 GEMM library that supports both...

DeepSeek (@deepseek_ai) — Wed, 26 Feb 2025 01:00:48 GMT

🚀 Day 3 of#OpenSourceWeekk: DeepGEMM

Introducing DeepGEMM - an FP8 GEMM library that supports both dense and MoE GEMMs, powering V3/R1 training and inference.

⚡ Up to 1350+ FP8 TFLOPS on Hopper GPUs
✅ No heavy dependency, as clean as a tutorial
✅ Fully Just-In-Time compiled
✅ Core logic at ~300 lines - yet outperforms expert-tuned kernels across most matrix sizes
✅ Supports dense layout and two MoE layouts

🔗 GitHubgithub.com/deepseek-ai/De…pT

💬466🔄854❤️6371👀967521📊1124

⚡ Powered by xgo.ing

🚀 Day 2 of #OpenSourceWeek: DeepEP Excited to introduce DeepEP - the first open-source EP communic...

DeepSeek (@deepseek_ai) — Tue, 25 Feb 2025 02:24:10 GMT

🚀 Day 2 of#OpenSourceWeekk: DeepEP

Excited to introduce DeepEP - the first open-source EP communication library for MoE model training and inference.

✅ Efficient and optimized all-to-all communication
✅ Both intranode and internode support with NVLink and RDMA
✅ High-throughput kernels for training and inference prefilling
✅ Low-latency kernels for inference decoding
✅ Native FP8 dispatch support
✅ Flexible GPU resource control for computation-communication overlapping

🔗 GitHubgithub.com/deepseek-ai/De…gW

💬509🔄1044❤️8080👀1388692📊1387

⚡ Powered by xgo.ing

🚀 Day 1 of #OpenSourceWeek: FlashMLA Honored to share FlashMLA - our efficient MLA decoding kernel...

DeepSeek (@deepseek_ai) — Mon, 24 Feb 2025 01:34:20 GMT

🚀 Day 1 of#OpenSourceWeekk: FlashMLA

Honored to share FlashMLA - our efficient MLA decoding kernel for Hopper GPUs, optimized for variable-length sequences and now in production.

✅ BF16 support
✅ Paged KV cache (block size 64)
⚡ 3000 GB/s memory-bound & 580 TFLOPS compute-bound on H800

🔗 Explore on GitHubgithub.com/deepseek-ai/Fl…X2

💬549🔄1306❤️10074👀1712716📊1687

⚡ Powered by xgo.ing

🚀 Day 0: Warming up for #OpenSourceWeek! We're a tiny team @deepseek_ai exploring AGI. Starting n...

DeepSeek (@deepseek_ai) — Fri, 21 Feb 2025 04:00:55 GMT

🚀 Day 0: Warming up for#OpenSourceWeekk! 

We're a tiny team @deepseek_ai exploring AGI. Starting next week, we'll be open-sourcing 5 repos, sharing our small but sincere progress with full transparency.

These humble building blocks in our online service have been documented, deployed and battle-tested in production.

As part of the open-source community, we believe that every line shared becomes collective momentum that accelerates the journey.

Daily unlocks are coming soon. No ivory towers - just pure garage-energy and community-driven innovation.

💬1379🔄2556❤️20270👀2544042📊3474

⚡ Powered by xgo.ing

🚀 Introducing NSA: A Hardware-Aligned and Natively Trainable Sparse Attention mechanism for ultra-f...

DeepSeek (@deepseek_ai) — Tue, 18 Feb 2025 07:04:05 GMT

🚀 Introducing NSA: A Hardware-Aligned and Natively Trainable Sparse Attention mechanism for ultra-fast long-context training & inference!

Core components of NSA:
• Dynamic hierarchical sparse strategy
• Coarse-grained token compression
• Fine-grained token selection

💡 With optimized design for modern hardware, NSA speeds up inference while reducing pre-training costs—without compromising performance. It matches or outperforms Full Attention models on general benchmarks, long-context tasks, and instruction-based reasoning.

📖 For more details, check out our paper herarxiv.org/abs/2502.11089UV7

💬879🔄2060❤️15178👀2612769📊2909

⚡ Powered by xgo.ing

🎉 Excited to see everyone’s enthusiasm for deploying DeepSeek-R1! Here are our recommended settings...

DeepSeek (@deepseek_ai) — Fri, 14 Feb 2025 08:56:47 GMT

🎉 Excited to see everyone’s enthusiasm for deploying DeepSeek-R1! Here are our recommended settings for the best experience:

• No system prompt
• Temperature: 0.6
• Official prompts for search & file upload: bit.ly/4hyH8npQ
• Guidelines to mitigate model bypass thinking:bit.ly/4gJrhkFr

The official DeepSeek deployment runs the same model as the open-source version—enjoy the full DeepSeek-R1 experience! 🚀

💬680🔄1655❤️15696👀1833080📊2860

⚡ Powered by xgo.ing

🚀 DeepSeek-R1 is here! ⚡ Performance on par with OpenAI-o1 📖 Fully open-source model & techni...

DeepSeek (@deepseek_ai) — Mon, 20 Jan 2025 12:29:30 GMT

🚀 DeepSeek-R1 is here!

⚡ Performance on par with OpenAI-o1
📖 Fully open-source model & technical report
🏆 MIT licensed: Distill & commercialize freely!

🌐 Website & API are live now! Try DeepThink at https://chat.deepseek.com
🐋 1/n

💬1949🔄6714❤️35196👀12548771📊7371

⚡ Powered by xgo.ing

🎉 Introducing DeepSeek App! 💡 Powered by world-class DeepSeek-V3 🆓 FREE to use with seamless int...

DeepSeek (@deepseek_ai) — Wed, 15 Jan 2025 09:47:47 GMT

🎉 Introducing DeepSeek App!

💡 Powered by world-class DeepSeek-V3
🆓 FREE to use with seamless interaction
📱 Now officially available on App Store & Google Play & Major Android markets
🔗Download now: httdownload.deepseek.com/app/� 1/3

💬312🔄730❤️4058👀1378374📊839

⚡ Powered by xgo.ing

DeepSeek has not issued any cryptocurrency. Currently, there is only one official account on the Twi...

DeepSeek (@deepseek_ai) — Fri, 10 Jan 2025 10:27:46 GMT

DeepSeek has not issued any cryptocurrency. Currently, there is only one official account on the Twitter platform. We will not contact anyone through other accounts.Please stay vigilant and guard against potential scams.

💬594🔄659❤️6351👀2218978📊1120

⚡ Powered by xgo.ing

🚀 Introducing DeepSeek-V3! Biggest leap forward yet: ⚡ 60 tokens/second (3x faster than V2!) 💪 En...

DeepSeek (@deepseek_ai) — Thu, 26 Dec 2024 11:26:48 GMT

🚀 Introducing DeepSeek-V3!

Biggest leap forward yet:
⚡ 60 tokens/second (3x faster than V2!)
💪 Enhanced capabilities
🛠 API compatibility intact
🌍 Fully open-source models & papers

🐋 1/n

💬646🔄2098❤️12972👀7393318📊2732

⚡ Powered by xgo.ing

🚀 DeepSeek-R1-Lite-Preview is now live: unleashing supercharged reasoning power! 🔍 o1-preview-lev...

DeepSeek (@deepseek_ai) — Wed, 20 Nov 2024 11:40:30 GMT

🚀 DeepSeek-R1-Lite-Preview is now live: unleashing supercharged reasoning power!

🔍 o1-preview-level performance on AIME & MATH benchmarks.
💡 Transparent thought process in real-time.
🛠️ Open-source models & API coming soon!

🌐 Try it now at httchat.deepseek.comeepSeek

💬257🔄773❤️4396👀3216641📊1051