@ArizePhoenix: For agents, active parameters are what you pay for in latency and cost, and agent loops resend context, retry tool call…
Summary
Arize Phoenix introduces M3, which adds sparse attention and a 1M-token context window to reduce latency and cost in AI agents by keeping long tool histories in context.
Similar Articles
Your agent isn't expensive. Your context window is. Here's the math
Explains why agent API bills grow quadratically with context length because each turn re-reads the full history, and shares practical techniques like expiring tool results, shrinking tool schemas, and compacting context to cut costs.
MiniMax M3: How Sparse Attention Makes Long-Horizon Agents Practical (11 minute read)
MiniMax M3 uses sparse attention to make long-horizon agents practical by keeping context cost predictable and low, enabling 500K-token contexts with minimal quality loss and significant speedups in production.
How are you handling active context once durable agent memory actually works?
The article discusses the challenges of managing active context in long-running AI agent workflows, focusing on balancing context retention with efficiency and cost, and seeks practical solutions from the community.
@rohanpaul_ai: Long-running agents do not just need a bigger context window. They need to learn what deserves to stay in context at al…
ContextPilot teaches AI agents to manage context proactively via fine-grained reinforcement learning, improving performance on long-context benchmarks by focusing on what information to retain.
@lateinteraction: Agents often externalize some context: a repository in coding agents, a corpus in RAG, and the user prompt in an RLM. N…
New research by Joshua Gu shows that AI agents perform better when they manage a small buffer in their context window as a cache for external context, challenging the common practice of pushing context entirely out of the prompt.