HiVe-Mem cuts multimodal agent memory tokens by about 90%
A graph of textual and visual anchors plus a trainable evidence selector improved three memory benchmarks across base models.
Stories per day · 30 days
2 RSS
Relatedmemory2retrieval2agent-benchmarks1claude1context-management1language-models1lifelong-agents1llm-agents1long-horizon1multimodal-agents1osworld1reasoning-budget1task-composition1
16 stories, newest first · one card per story
Stories tagged #token-efficiency
TokenWorth scores memory worth for lifelong agents under tight budgets
Across three LifelongAgentBench settings it records the highest success while emitting the fewest output tokens of the methods tested.
Language models err more on composed tasks via implicit token budgets
Across open-weight models and Claude Opus 5, grouping fragile tasks cut tokens 43% without losing accuracy.
CarryOn organizes long-horizon agent context around task objects
Authors report gains of 5.8% and 7.0% with token cuts of 51.9% and 22.9% on two benchmarks.
Branching On-Policy Distillation
BranchOPD branches uncertain on-policy rollouts, beating vanilla OPD by 18.55 math and 13.8 ALFWorld points.
VisionWeave: Weaving Elastic Visual Representations as a Native Capability of MLLMs
VisionWeave cuts MLLM visual tokens 43% on Qwen3.8-27B while retaining 98.9% performance and 2.3x SGLang throughput.
Fewer Tokens, Better Action: GPT-6 Astra Robot Agents with 14% Higher Success Rate but 65% Fewer Tokens
PyRUA-Lean raises GPT-6 Astra robot-agent success from 63.1% to 71.7% while cutting input tokens 65%.
Synthetic Pre-pretraining Survives Scale, but Not as a Grammatical Prior
Synthetic pre-pretraining still saves tokens up to 7B parameters, but gains come from retrieval, not grammar.
Fireworks AI Releases Ember-1: A Post-Trained Kimi K3 That Uses About 40% Fewer Tokens
Fireworks AI launched Ember-1, a post-trained Kimi K3 that cuts reasoning tokens ~40% at matched accuracy, API-only.
3
Nvidia's SoL-Pi system cuts coding agent token usage nearly in half by optimizing the harness
Nvidia's SoL-Pi research system cuts coding-agent token use nearly in half by optimizing the harness.
Capable yet Parsimonious: Extracting and Characterizing Hidden Chain-of-Thought in Frontier Models
Researchers extract hidden chain-of-thought from frontier models, including GPT-6 Astra, and compare reasoning styles.
SoL-Pi: Recursively Scaling Auto-Research Loops for Efficient Agent Harness
SoL-Pi scales auto-research loops to produce an agent harness matching Pi while cutting token traffic 44.7-49.0% and API cost by a third.
Region-Level Policy Optimization for Fine-grained MLLM Perception
Vision-RL2 applies region-level reinforcement learning to MLLM perception, surpassing largest-budget accuracy with roughly four times fewer visual tokens.
Ask HN: How do you manage skills files?
A Hacker News thread debates whether agent skill files are worth managing, citing 2–4x output-token reductions on flagship models in one company's testing.
4
Introducing agentic video understanding with Gemini
Google DeepMind launches agentic video understanding for Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite, cutting video-analysis tokens up to 88%.
8
Thinking of ACE? We Can Do It with Fewer Tokens
IBM Research's ALTK-EVOLVE-SLDD blog post claims ACE-style LLM analysis using substantially fewer tokens.