Sign InOpen Brain
Atlas / Context

Context Caching

Open JSONConfidence: Auto-collectedLast updated 2026-07-31

Current Answer

No editorial synthesis yet — the evidence below is collected automatically from source labels. A current answer lands here once an editor approves one.

Evidence

GoogleDocs UpdateOfficial Source
Gemini agent SDK adds shared context caching across sessions
Cached context is billed once and reused across agent sessions — changes the economics of long system prompts.
TokTier: Exact Stateful Tokenization for Agentic LLM Serving
arXiv · 2026-07-31

TokTier targets a hidden agent-serving bottleneck: repeatedly tokenizing long transcripts after small tool results. Its stateful repair cut median vLLM time to first token by 16–34% in reported tests.

ReToken: One Token to Improve Vision-Language Models for Visual Retrieval
arXiv · 2026-07-30

ReToken adds one learned retrieval embedding to select relevant visual tokens from a cached context, improving long-image and video benchmarks while fitting training and inference on one H100.

How enabling two settings tripled our scores on the ARC-AGI-3 benchmark
OpenAI · 2026-07-29

Two API settings—reasoning retention and compaction—reportedly tripled GPT-5.6’s ARC-AGI-3 score. Agent evals should treat runtime configuration as part of the tested system.

Eviction as Estimation: A Fixed-Lag Smoothing View of Test-Time Memory, and When Measuring Beats Accumulating
arXiv · 2026-07-27

A delayed KV-eviction policy helps when memory reuse is sharp and observable, but loses its edge on natural-text benchmarks—a useful warning for agent-memory claims.

diegosouzapw/OmniRoute
GitHub

OmniRoute puts many model providers behind one OpenAI-compatible endpoint, with routing, quota failover, cost telemetry, compression, and coding-agent setup. Its breadth raises operational and trust questions.

esengine/DeepSeek-Reasonix
GitHub

Reasonix is a configurable terminal coding agent built to preserve DeepSeek prefix-cache reuse across long sessions, with optional planner/executor models and subprocess-based plugins.

Stable permalink · evidence auto-collected from source labels · synthesis maintained by feed7 editorial