Topic · Concept
Long context
Long context refers to a model's ability to take in very large inputs, now commonly a million tokens or more, in a single request. Bigger windows let a model read whole codebases or long documents at once, but a large window is not the same as reliable recall of everything in it.
Fact file
"Lost in the Middle" findingRecall is often best at the start or end of the context (Liu et al., Jul 6, 2023)
Gemini 1.5 research claim>99% retrieval up to at least 10M tokens (vendor research, Mar 8, 2024)
Start here
Latest on Long context
Oct 9Open source · NewsDeepSeek-V4.1-Flash: an MIT-licensed open-weights MoE that DeepSeek says beats its own flagshipOct 9Models · BenchmarksLong-context recall: what independent benchmarks say "1 million tokens" really buys youOct 9Apps · CompareChatGPT vs Gemini for research: deep research, long documents, prices and privacy comparedOct 9Apps · ExplainerHow Gemini handles a 900-page PDF: the limits, the costs and what long-context benchmarks showOct 9Research · ExplainerContext windows vs memory vs RAG: what each one actually does
Nothing in this format on this page — try another filter or the next page.
About this hub
Long context refers to a model's ability to take in very large inputs, now commonly a million tokens or more, in a…