Subquadratic launches with a $29M seed and debuts SubQ, an LLM that uses a subquadratic sparse attention architecture to achieve a 12M-token context window
Subquadratic, a company developing a novel generative artificial intelligence model, launched today with $29 million in seed funding.
SiliconANGLEKyt Dotson
Context & Ripple Effects
Subquadratic enters a field already focused on making language models more useful over large bodies of information. Related coverage includes Alibaba’s Qwen3-Next, which was also positioned around long-context understanding and computational efficiency, while Contextual AI and Poetiq reflect parallel efforts to tailor LLM capabilities to enterprise and task-specific use cases.
The launch pairs an architectural claim—a 12M-token context window enabled by subquadratic sparse attention—with seed financing, giving Subquadratic resources to turn that technical position into a product and developer proposition.
First-order effects
Subquadratic gains $29M in seed capital and launches SubQ, immediately establishing the company as a long-context LLM contender.
Potential users evaluating workloads that require very large input corpora now have another model architecture to assess against existing long-context offerings.
Second-order effects
Long-context model vendors face added pressure to demonstrate not only maximum context size, but the practical quality and efficiency of handling that context.
The competitive focus shifts toward architecture-level differentiation: sparse-attention approaches can become a more central comparison point for buyers and developers than model scale alone.
Third-order effects
If long-context systems continue to improve through more efficient attention designs, context capacity may become a more accessible product dimension rather than a feature limited to the largest model platforms.
That would favor an LLM market differentiated by workload-specific architecture and deployment economics, though the durability of Subquadratic’s advantage will depend on real-world performance beyond its stated context window.
The trend: This is one data point in the push to expand usable LLM context through architectural efficiency rather than relying solely on ever-larger models and compute budgets.
Introducing SubQ - a major breakthrough in LLM intelligence. It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA), And the first frontier model with a 12 million token context window which is: - 52x faster than FlashAttention at 1MM tokens - [v…
The numbers behind the SubQ announcement: Speed: 52x faster than Flash Attention SWE Bench Verified: 81.8% Ruler (128K): 95% MRCR V2: 65.9% Get early access at https://subq.ai/
With so much capital concentrating in so few private companies, and Anthropic and OpenAI breaking all “startup” growth norms, it's easy to forget that we are still incredibly early in this AI wave. On that note, I am THRILLED @subquadratic is now out of stealth. This is a
“early access” Scammy vibes If it's really a sub-quadratic sparse attention arch (SSA), then serving this should be really cheap No point in putting this behind early access
A 12-million-token context window at 1,000x less compute capable of coding for weeks at a time and filing hundreds of PRs in the process. 🤯 If you thought AI can run laps around us now, the rate of progress in the next few years will become exponential.
Let's read the technical report. TLDR; No real answers on how their method works. Doesn't make me feel better about it. They seem to understand the problem: “[Attention] is expensive for the same reason: every query compares against every key. The result is an all-pairs
SubQ is either the biggest breakthrough since the Transformer... > 52x faster than FlashAttention at 1mm tok context > 20x cheaper than Opus ...or it's AI Theranos. Requested early access so hopefully can investigate soon. [image]
does not pass my smell test.. > a breakthrough like this gets published at ICML/NeurIPS/ICLR - not with a startup launch video - would love to read a preprint atleast (technical report coming soon is v SUS) > usual suspects engagement boosting this tweet was the final straw
Very welcome to see more research in that space but a bit puzzling until the report clarifies: *Clearly a continuous pretrain of an open weight model (totally fair for this but we'll need a before and after). *No actually long evals (>1M) even though RULER could be extrapolated.
nevermind no longer trying to give them the benefit of the doubt here: they claimed O(n) and ‘subq is linear vs quadratic’ which is pretty ridiculous the speedup numbers in their announcement video don't seem to line up with this? and just 12M context with O(n) scaling? this is […
my first take, and a good lesson on good research epistemics here: what can we infer from ~82% SWE-Bench? it's possible they (1) they trained a new model, from scratch, that is unlike a regular transformer but i've never heard of this company before, and checking their funding [i…