Ollama's default context length is inconsistently documented and applied across different configurations. The documentation presents at least three conflicting default values: 2048 in the Modelfile reference, variable lengths (4k, 32k, 256k) based on available VRAM, and a value baked into the model image itself. The most specific setting, such as one defined in a model's Modelfile or in an individual request's options, overrides broader defaults like environment variables or VRAM-based tiers. A significant issue is that when a prompt exceeds the effective context window, Ollama silently drops earlier messages without any user-visible indication or error, logging the truncation only at a debug level. AI
IMPACT Users may encounter unexpected behavior with long conversations due to silent truncation of context, impacting the reliability of Ollama for complex tasks.
RANK_REASON Article discusses configuration and behavior of an existing AI tool, not a new release or significant industry event.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
