Cognee vs Minerva

Cognee turns documents, chats, tickets, and API data into graph memory language model agents can recall across sessions. It builds linked entity graphs from that context so agents stop restarting from scratch each run. You can connect Slack, GitHub, or Linear so coding agents and support bots share one permission-aware company brain.

Plain RAG returns similar text chunks. Cognee pairs vector search with knowledge graphs and auto-generated ontologies, so recall pulls connected entities and cited facts rather than isolated snippets. The SDK centers on four verbs, remember, recall, forget, and improve, and the same surface ships over HTTP and MCP for Claude Code, Codex, and OpenClaw.

Platform teams use it for coding agent memory, GraphRAG pipelines, deal intelligence, and customer-facing agents that need grounded answers. Run it with pip locally, self-host in Docker or on-prem, or move to Cognee Cloud when you want managed scale. The project reports 30.4k GitHub stars and 5M+ SDK runs per month, with production deployments at Bayer and Knowunity.

Minerva is a large language model from Google Research built to solve math and science questions through step-by-step written reasoning. It reads problems that mix plain English with LaTeX notation, then writes out solutions involving arithmetic, algebra, and symbolic steps. The model was trained on scientific papers and web pages where mathematical formatting was kept intact, rather than stripped during preprocessing.

Most math-capable models lean on external tools like Python interpreters or calculators at inference time. Minerva takes the opposite bet: it generates full worked solutions from the model weights alone, using chain-of-thought prompting and majority voting across multiple sampled answers. That informal approach covers a wider range of problem types than formal theorem provers, but the trade-off is answers cannot be machine-verified the way Coq or Lean proofs can.

Researchers studying quantitative reasoning in language models use Minerva as a reference point for STEM benchmark performance. The public sample explorer hosts 110 solved problems across algebra, physics, chemistry, and other topics, so anyone can read through how the model arrived at each answer. Educators and ML engineers reviewing benchmark methodology will find the published MATH, MMLU-STEM, GSM8k, and OCWCourses scores useful for comparing against newer models.