# SubQ

11 Hacker News commenters recommended SubQ and 14 criticized it, November 2025 to November 2026. It was recommended in 3 of 13 months.

**Verdict:** People recommend SubQ for claimed long-context reasoning that is faster and cheaper than transformer models. The most common caveat is the missing technical report and no independent verification.

## Why people recommend it

- Solves long context in architecture (3 comments): "Awesome stuff. Solving context at the model architecture layer rather than trying to bolt on extra memory is the right direction IMO." ([source](https://news.ycombinator.com/item?id=48556847)) "This is pretty remarkable. We've spent a lot of time finding workarounds for LLMs reading long docs." ([source](https://news.ycombinator.com/item?id=48023242)) "Looks like long context isn’t a problem anymore" ([source](https://news.ycombinator.com/item?id=48023665))
- Faster and cheaper than transformers (2 comments): "doesn’t have context rot, is orders of magnitude faster, and uses less compute that current transformer models" ([source](https://news.ycombinator.com/item?id=48024015)) "SubQ claims to reason as good as the frontier models, not better, but reasonably as good, and is 5x cheaper" ([source](https://news.ycombinator.com/item?id=48195788))
- Seen as architectural breakthrough (2 comments): "Awesome stuff. Solving context at the model architecture layer rather than trying to bolt on extra memory is the right direction IMO." ([source](https://news.ycombinator.com/item?id=48556847)) "it's a real architectural breakthrough if true" ([source](https://news.ycombinator.com/item?id=48024213))
- Impressive technology overall (2 comments): "This is pretty remarkable. We've spent a lot of time finding workarounds for LLMs reading long docs." ([source](https://news.ycombinator.com/item?id=48023242)) "I love the tech. It does amazing things." ([source](https://news.ycombinator.com/item?id=48558223))

## What people warn about

- Looks like hype or a scam (6 comments): "The only purpose of this press release seems to be making investors give them more money." ([source](https://news.ycombinator.com/item?id=48558736)) "It's been all talk and no action ever since their first announcement." ([source](https://news.ycombinator.com/item?id=48556992)) "Everything in south Florida is a scam, this is just one of many" ([source](https://news.ycombinator.com/item?id=48620338))
- No technical report or details (5 comments): "Can you back up your claims? Why did you not release the white paper in parallel with the product? Feels really fishy." ([source](https://news.ycombinator.com/item?id=48038926)) "The mechanism by which SSA meets these requirements is outside the scope of this report" ([source](https://news.ycombinator.com/item?id=48561683)) "Disappointing they don't actually say how their sparse attention mechanism works." ([source](https://news.ycombinator.com/item?id=48556528))
- Lacks transparency, no independent verification (3 comments): "it seems that they aren't being fully transparent" ([source](https://news.ycombinator.com/item?id=48557153)) "Can you back up your claims? Why did you not release the white paper in parallel with the product? Feels really fishy." ([source](https://news.ycombinator.com/item?id=48038926)) "No API access for independent verification - vaporware." ([source](https://news.ycombinator.com/item?id=48076199))
- Quality lags frontier models (2 comments): "The delta between their score (81%) vs. Opus (87%) on SWE bench was hugely minimized" ([source](https://news.ycombinator.com/item?id=48029016)) "12 million token context with worse quality means this isn't going anywhere" ([source](https://news.ycombinator.com/item?id=48094325))

## Alternatives hackers recommend

- [Claude Opus](https://hackerapproved.com/p/claude-opus.md): 3331 recommend
- [OpenAI GPT](https://hackerapproved.com/p/gpt-5-5.md): 1205 recommend
- [Claude Sonnet](https://hackerapproved.com/p/claude-sonnet.md): 854 recommend
- [DeepSeek V4 Flash](https://hackerapproved.com/p/deepseek-v4-flash.md): 828 recommend
- [Fable](https://hackerapproved.com/p/fable.md): 823 recommend

## Cite as

Hacker Approved, "SubQ", https://hackerapproved.com/p/subq/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).