SubQ – a sub-quadratic LLM built for multi-million token reasoning

Subquadratic launched SubQ, a sub-quadratic sparse-attention LLM with a 12-million-token context window, enabling multi-million token reasoning at linear cost. The model reduces attention compute by nearly 1,000× at 12M tokens and outperforms GPT-5.5 and Opus 4.8 on benchmarks like LiveCodeBench v6 and GPQA Diamond.

API For developers and teamsThe full-context API for developers and enterprise teams. Process full repositories and pipeline states in a single API call at linear cost. - → 12M token context window - → Streaming + tool use - → OpenAI-compatible endpoints SubQ is a sub-quadratic LLM built for multi-million token reasoning , allowing agents to work across full repositories, long histories, and persistent state without quality loss. Use Cases Reason across millions of tokens in one prompt: entire repos, whole artifacts, and long-running agent state, with room to spare at a fraction of the cost . ~ Approximate token counts. Architecture SubQ is the first model built on a fully sub-quadratic sparse-attention architecture. LLMs today waste compute by processing every possible relationship between words, but only a small fraction of these relationships matter. SubQ finds and focuses only on those, ensuring compute is used where it matters most. At 12M tokens, this reduces attention compute almost 1,000×, changing the way LLMs scale. Benchmarks SubQ has near-perfect performance on single-fact retrieval and multi-task retrieval, both at scale. SubQ balances long-context retrieval without compromising on reasoning and knowledge. | Benchmark | SubQ 1.1 Small | GPT-5.5 | Opus 4.8 | Sonnet 4.6 | GPT-5.4-mini | GPT-5.4-nano | Haiku 4.5 | |---|---|---|---|---|---|---|---| Graduate-level science GPQA Diamond · pass@1 | 85.4 | 93.2 | 92 | 87.5 | 87.5 | 81.7 | 67.2 | Agentic finance AutomationBench | 13% | 18% | 16% | 8% | 0% | n/r | 3% | Competitive programming LiveCodeBench v6 · pass@4 | 89.7 | 92 | 92.2 | 88.9 | 78.6 | 78.2 | 69.7 | SubQ uses 64.5x less compute than dense attention, and is 56× faster than FlashAttention-2 at 1M-token context. Products The full-context API for developers and enterprise teams. Process full repositories and pipeline states in a single API call at linear cost. The long-context layer for coding agents. Plug into Claude Code, Codex, and Cursor to map codebases, gather context, and answer token-heavy questions faster. About Subquadratic is a frontier AI research and infrastructure company building a new class of LLMs. While other major labs focus on incremental improvements to Transformer models, we're pushing foundational change at the model architecture level — enabling large-context, multi-modal inference that scales efficiently where transformers can't. Built by researchers from Early Access Join the private preview.