# DeepSeek R1

25 Hacker News commenters recommended DeepSeek R1 and 15 criticized it, November 2025 to November 2026. It was recommended in 10 of 13 months.

**Verdict:** Free open-weight reasoning model under an MIT license that people run locally and find capable. Critics point to language mixing, loops, and censorship.

## Why people recommend it

- Runs locally, even on modest hardware (5 comments): "The smallest Deepseek R1 8B, running locally on CPU only, casually mentioning Efinix Trion FPGA fabrics" ([source](https://news.ycombinator.com/item?id=48408215)) "I have my best results with Ollama using deepseek-r1." ([source](https://news.ycombinator.com/item?id=47484086)) "You could also try the 14B Deepseek R1." ([source](https://news.ycombinator.com/item?id=48509556))
- Open weights under MIT license (4 comments): "it was pretty good compared with OpenAI, but also fully self-hostable" ([source](https://news.ycombinator.com/item?id=47064939)) "the weights are actually under a real open source license, (MIT)" ([source](https://news.ycombinator.com/item?id=47888116)) "Once you have the model, you have the model." ([source](https://news.ycombinator.com/item?id=48693262))
- Groundbreaking and influential releases (3 comments): "DeepSeek R1 and DeepSeekMath were groundbreaking models which were substantially stronger than any models available for distillation at the time they were developed" ([source](https://news.ycombinator.com/item?id=50217516)) "All their findings in R1 are highly plausible and most are replicated to some degree and adopted in the research and industry." ([source](https://news.ycombinator.com/item?id=46188534)) "Every major release from them has brought something really new to the field" ([source](https://news.ycombinator.com/item?id=48586396))
- Capable, convincing answers (3 comments): "it just did it -- not just decently, but with less slop than Claude/OpenAI" ([source](https://news.ycombinator.com/item?id=47167419)) "Deepseek R1's sounds pretty convincing to me, but I'm not a good judge." ([source](https://news.ycombinator.com/item?id=47733749)) "Deepseek R1 was also capable of giving a correct response." ([source](https://news.ycombinator.com/item?id=49564136))

## What people warn about

- Censorship and bias in answers (2 comments): "the local model - full-weight 685B, not the distilled ones - was pretty much unhinged regarding censorship on any topic" ([source](https://news.ycombinator.com/item?id=47681977)) "recent deepseek models are not nearly as biased as R1" ([source](https://news.ycombinator.com/item?id=50256060))
- Weaker than larger models on benchmarks (2 comments): "even the 122b is more capable than Deepseek R1 (a 671b model) on most benchmarks" ([source](https://news.ycombinator.com/item?id=48623127)) "The benchmarks lie. Go try coding full-time with R1 vs Codex or GPT-5" ([source](https://news.ycombinator.com/item?id=45863816))
- Severe security vulnerabilities in code (2 comments): "the likelihood of it generating code with severe vulnerabilities increased to 27.2%" ([source](https://news.ycombinator.com/item?id=50280036)) "political context alone determined whether basic security controls existed" ([source](https://news.ycombinator.com/item?id=46113688))

## Alternatives hackers recommend

- [OpenAI o3](https://hackerapproved.com/p/o3.md): 37 recommend
- [VibeThinker 3B](https://hackerapproved.com/p/vibethinker-3b.md): 28 recommend
- [OpenAI o1](https://hackerapproved.com/p/openai-o1.md): 12 recommend

## Cite as

Hacker Approved, "DeepSeek R1", https://hackerapproved.com/p/deepseek-r1/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).