model
Large language model from Chinese AI developer DeepSeek, preceding the V4 release and used as a benchmark for context-length and efficiency comparisons.
No evidence-backed relationships are recorded.
A 73-page preprint evolved instructions that jumped between coding agents and wrote themselves into the file that becomes the next system prompt. A short warning nearly stopped transmission.
Perspective Coverage
4 publishersReality
The efficiency claims are self-reported and measured against DeepSeek's own prior model, but the weights are MIT-licensed and already pulled 1.78 million times, which is what turns a ratio into a number a buyer can carry into a renewal.
Reality
OpenRouter says DeepSeek went from 9% to 18% of tokens in six months, with agent loops doing most of the work. Token share and spend share are not the same league table.
Reality
Locally Uncensored put four runs of one bugfix on the same model at the same prices. Its own 2.6.5 agent cost 4395 credits, about 3.4 times its successor and double opencode's average.
Reality
An Anthropic and EPFL preprint shows plain-language goals hopping agent to agent through persistent files, and a one-paragraph warning in the system prompt stopping nearly all of it.
Reality
IBM Research ran self-mined guidelines across eight models on AppWorld. One model gained 16.1 points for 5 percent more tokens; another gained nothing at all.
Reality
Self-propagating payloads did move between agents through editable soul files, but one inoculation paragraph held against 150-plus optimized strains, and nothing propagated in the wild.
Reality