# \#deepseek

**URL:** https://forum.devtalk.com/tag/deepseek/3148.md

[Latest](https://forum.devtalk.com/latest.md) · [Categories](https://forum.devtalk.com/categories.md) · [Tags](https://forum.devtalk.com/tags.md)

---

## [DeepSeek V4—almost on the frontier, a fraction of the price](https://forum.devtalk.com/t/deepseek-v4-almost-on-the-frontier-a-fraction-of-the-price/243615)

<div class="topic-metadata">

**Author:** [@CommunityNews](https://forum.devtalk.com/u/CommunityNews)\
**Replies:** 7\
**Last updated:** [5 October 2026 16:50 UTC](https://forum.devtalk.com/t/deepseek-v4-almost-on-the-frontier-a-fraction-of-the-price/243615 "2026-10-05T16:50:06Z")

</div>

Chinese AI lab DeepSeek’s last model release was V3.2 (and V3.2 Speciale) last December. They just dropped the first of their hotly anticipated V4 series in the shape of two … Read in full here:

---

## [Reasonix — DeepSeek-native AI coding agent](https://forum.devtalk.com/t/reasonix-deepseek-native-ai-coding-agent/246070)

<div class="topic-metadata">

**Author:** [@CommunityNews](https://forum.devtalk.com/u/CommunityNews)\
**Replies:** 1\
**Last updated:** [22 June 2026 17:11 UTC](https://forum.devtalk.com/t/reasonix-deepseek-native-ai-coding-agent/246070 "2026-06-22T17:11:22Z")

</div>

Open-source AI coding agent for your terminal. Engineered around DeepSeek’s prefix-cache. MCP first-class · plan mode · MIT. Read in full here:

---

## [DeepSeek V4 is live in preview — should your team switch?](https://forum.devtalk.com/t/deepseek-v4-is-live-in-preview-should-your-team-switch/242682)

<div class="topic-metadata">

**Author:** [@xiji2646-netizen](https://forum.devtalk.com/u/xiji2646-netizen)\
**Replies:** 0\
**Last updated:** [25 April 2026 20:33 UTC](https://forum.devtalk.com/t/deepseek-v4-is-live-in-preview-should-your-team-switch/242682 "2026-04-25T20:33:47Z")

</div>

DeepSeek officially launched deepseek-v4-flash and deepseek-v4-pro in preview on April 24, 2026. The legacy routes (deepseek-chat, deepseek-reasoner) are deprecated July 24, 2026. I’ve been thinking through how to actua…

---

## [DeepSeek V4 dropped today — $0.28/M output on 1M context, running on Huawei Ascend. Are you routing workloads to it?](https://forum.devtalk.com/t/deepseek-v4-dropped-today-0-28-m-output-on-1m-context-running-on-huawei-ascend-are-you-routing-workloads-to-it/242498)

<div class="topic-metadata">

**Author:** [@xiji2646-netizen](https://forum.devtalk.com/u/xiji2646-netizen)\
**Replies:** 1\
**Last updated:** [25 April 2026 00:18 UTC](https://forum.devtalk.com/t/deepseek-v4-dropped-today-0-28-m-output-on-1m-context-running-on-huawei-ascend-are-you-routing-workloads-to-it/242498 "2026-04-25T00:18:04Z")

</div>

DeepSeek just released V4 and the pricing is hard to ignore. V4-Flash: $0.28/M output tokens. V4-Pro: $2.19/M. Both with 1M token context as default. For reference: GPT-4 Turbo is $30/M output. Claude Opus 4.6 is $75/M…

---

## [Build a DeepSeek Model (From Scratch) (Manning)](https://forum.devtalk.com/t/build-a-deepseek-model-from-scratch-manning/221495)

<div class="topic-metadata">

**Author:** [@ManningBooks](https://forum.devtalk.com/u/ManningBooks)\
**Replies:** 3\
**Last updated:** [28 November 2025 21:20 UTC](https://forum.devtalk.com/t/build-a-deepseek-model-from-scratch-manning/221495 "2025-11-28T21:20:19Z")

</div>

In Build a DeepSeek Model (From Scratch) you’ll build your own DeepSeek clone from the ground up. First, you’ll quickly review LLM fundamentals, with an eye to where DeepSeek’s innovations address the common problems an…

---

## [How to: Run DeepSeek on Mac, Windows, and Linux!](https://forum.devtalk.com/t/how-to-run-deepseek-on-mac-windows-and-linux/185713)

<div class="topic-metadata">

**Author:** [@AstonJ](https://forum.devtalk.com/u/AstonJ)\
**Replies:** 10\
**Last updated:** [6 February 2025 06:44 UTC](https://forum.devtalk.com/t/how-to-run-deepseek-on-mac-windows-and-linux/185713 "2025-02-06T06:44:03Z")

</div>

This is a very quick guide, you just need to: Download LM Studio: https://lmstudio.ai/ Click on search Type DeepSeek, then select the one you want\* and click download It should then show in your models \*When deciding …

---

## [DeepSeek is a Game Changer for AI (video by Computerphile)](https://forum.devtalk.com/t/deepseek-is-a-game-changer-for-ai-video-by-computerphile/185734)

<div class="topic-metadata">

**Author:** [@AstonJ](https://forum.devtalk.com/u/AstonJ)\
**Replies:** 0\
**Last updated:** [29 January 2025 20:45 UTC](https://forum.devtalk.com/t/deepseek-is-a-game-changer-for-ai-video-by-computerphile/185734 "2025-01-29T20:45:13Z")

</div>

This is probably one of the best technical (yet accessible) breakdowns around atm:

---

## [DeepSeek-V3 Technical Report](https://forum.devtalk.com/t/deepseek-v3-technical-report/192779)

<div class="topic-metadata">

**Author:** [@CommunityNews](https://forum.devtalk.com/u/CommunityNews)\
**Replies:** 0\
**Last updated:** [27 March 2025 14:46 UTC](https://forum.devtalk.com/t/deepseek-v3-technical-report/192779 "2025-03-27T14:46:32Z")

</div>

We present DeepSeek-V3, a strong Mixture-of-Experts (MoE) language model with 671B total parameters with 37B activated for each token. To achieve efficient inference and cost-effective training, DeepSeek-V3 adopts Multi…

---

## [How to run Ollama deepseek-coder:6.7b-instruct-q4\_K\_M in Docker for CrewAI Agents?](https://forum.devtalk.com/t/how-to-run-ollama-deepseek-coder-6-7b-instruct-q4-k-m-in-docker-for-crewai-agents/190843)

<div class="topic-metadata">

**Author:** [@yamuna](https://forum.devtalk.com/u/yamuna)\
**Replies:** 0\
**Last updated:** [12 March 2025 08:29 UTC](https://forum.devtalk.com/t/how-to-run-ollama-deepseek-coder-6-7b-instruct-q4-k-m-in-docker-for-crewai-agents/190843 "2025-03-12T08:29:11Z")

</div>

Hi everyone, I’m trying to run deepseek-coder:6.7b-instruct-q4\_K\_M in Docker using Ollama to create an LLM that will be used by CrewAI agents. However, I’m encountering errors while initializing the SQL Query Generator …

---

## [Post your DeepSeek results](https://forum.devtalk.com/t/post-your-deepseek-results/185733)

<div class="topic-metadata">

**Author:** [@AstonJ](https://forum.devtalk.com/u/AstonJ)\
**Replies:** 15\
**Last updated:** [6 March 2025 23:29 UTC](https://forum.devtalk.com/t/post-your-deepseek-results/185733 "2025-03-06T23:29:12Z")

</div>

Curious what kind of results others are getting, I think actually prefer the 7B model to the 32B model, not only is it faster but the quality of results seem better too. Make we could use this thread to post prompts and…

---

## [Is it possible to build an online shop recommendation system with Deepseek?](https://forum.devtalk.com/t/is-it-possible-to-build-an-online-shop-recommendation-system-with-deepseek/187291)

<div class="topic-metadata">

**Author:** [@trazorx](https://forum.devtalk.com/u/trazorx)\
**Replies:** 0\
**Last updated:** [12 February 2025 12:36 UTC](https://forum.devtalk.com/t/is-it-possible-to-build-an-online-shop-recommendation-system-with-deepseek/187291 "2025-02-12T12:36:45Z")

</div>

Is it possible to build a recommendation system for my online shop? I have online shop and i want to add a recommendation system on it and I’m wondering if I can do it by leveraging Deepseek. Anyone can help give any r…

---

## [DeepSeek - the free, open source “ChatGPT killer”](https://forum.devtalk.com/t/deepseek-the-free-open-source-chatgpt-killer/185588)

<div class="topic-metadata">

**Author:** [@AstonJ](https://forum.devtalk.com/u/AstonJ)\
**Replies:** 5\
**Last updated:** [3 February 2025 22:26 UTC](https://forum.devtalk.com/t/deepseek-the-free-open-source-chatgpt-killer/185588 "2025-02-03T22:26:29Z")

</div>

Loads of news stories about DeepSeek here in the last few days, no surprise as it’s been making headlines across the world! Currently a huge story on social media too. Anyone tried it? What do you think? Here’s a good …

---

## [DeepSeek (671B) running on a cluster of 8 Mac Mini Pros with 64GB RAM each](https://forum.devtalk.com/t/deepseek-671b-running-on-a-cluster-of-8-mac-mini-pros-with-64gb-ram-each/185709)

<div class="topic-metadata">

**Author:** [@AstonJ](https://forum.devtalk.com/u/AstonJ)\
**Replies:** 1\
**Last updated:** [29 January 2025 18:38 UTC](https://forum.devtalk.com/t/deepseek-671b-running-on-a-cluster-of-8-mac-mini-pros-with-64gb-ram-each/185709 "2025-01-29T18:38:13Z")

</div>

This is cool! DEEPSEEK-V3 ON M4 MAC: BLAZING FAST INFERENCE ON APPLE SILICON We just witnessed something incredible: the largest open-source language model flexing its muscles on Apple Silicon. We’re talking about the …

---

## [DeepSeek portal on Devtalk is now up](https://forum.devtalk.com/t/deepseek-portal-on-devtalk-is-now-up/185711)

<div class="topic-metadata">

**Author:** [@AstonJ](https://forum.devtalk.com/u/AstonJ)\
**Replies:** 0\
**Last updated:** [29 January 2025 13:45 UTC](https://forum.devtalk.com/t/deepseek-portal-on-devtalk-is-now-up/185711 "2025-01-29T13:45:22Z")

</div>

I have a feeling we’re going to see a lot of threads about DeepSeek, so have put up a portal for it :003:
