The Agent Architect #44: World-Model Simulation Planning

今週のパターン

World-Model Simulation Planning

概要:

エージェントは LLM や学習済みモデルで候補となる行動の結果をシミュレートし、それらを採点したうえで、最良のものを実環境で実行します。

使いどころ:

本番のウェブサイトや GUI 上での不可逆な操作で、後戻りが不可能または高コストであり、ミスの代償が大きい場面。

注意点:

シミュレーションの誤差は積み重なります。ロールアウトの早い段階での予測ミスが連鎖し、採点ロジックが正しくても行動選択を誤らせます。

今週のエージェントAI

• Serving DeepSeek-V4.1-Flash at Scale: vLLM Matches DeepInfer and Scales to GB300 NVL72vLLM

  With P/D disaggregation and Wide-EP on a GB300 NVL72 rack, vLLM serves DeepSeek-V4.1-Flash on SemiAnalysis AgentX at 954K tokens/s per GPU at a P90 of 196 tokens/s per user, 7.9x our aggregated single-node result.

• Questioning Instrumental Convergence in RLVR: Checkpoint Evidence from OLMo-3arXiv cs.AI

  Instrumental convergence is often invoked as a general reason to expect capable agents to acquire or express power-seeking subgoals: self-preservation, resource acquisition, deception, or resistance to oversight which poses an important threat to the society. For LLMs, reinforcement learning is the first stage of post-training that directly optimizes behavior against an objective; therefore, it is the stage where convergent instrumental behavior, on IC grounds, likely to appear. Does reinforcement learning with verifiable rewards (RLVR) make a model more likely to express instrumental-converge

• Voxtral 3Mistral AI

  Voxtral 3: Mistral's voice agent SDK for real-time speech in 50+ languages — accurate transcription, emotional TTS, deep tool use. EU-hosted or self-deployed.

• GLM-5.3 Optimizations, Part 2: PCP, DCP, and Hybrid HiSparse on 16 H200svLLM

  vLLM serves GLM-5.3 on 16 H200s with PCP and DCP prefill and Hybrid HiSparse decode, reaching 1M-token contexts and 1.7–3.2× the per-user speed of data-parallel P/D on the AgentX agentic benchmark.

• From Reasoning to Action: Benchmarking Frontier AI Models on LIBERO-10 Robot Manipulation TasksarXiv cs.AI

  Recent advances in frontier AI models have demonstrated remarkable capabilities in visual understanding, reasoning, planning, and tool use, raising the possibility of using general-purpose AI models for robot manipulation without task-specific training. However, their manipulation capabilities have not been systematically evaluated on standardized robotics benchmarks. In this report, we present a unified closed-loop manipulation framework for evaluating frontier AI models on the LIBERO-10 benchmark. Our framework enables models to control a robot directly through visual observations, robot pro