最新模型

新模型、開放權重、多模態、語音、影像與模型架構。

MoNe Writes Long Context into Test-Time Neural Memory, Reducing Total FLOPs for 128K Inference by 81%

Qualcomm AI Research adds per-layer fast weights to a frozen Transformer, absorbing context in 512-token segments so queries no longer need to read the source text. In 128K experiments, Qwen2.5-0.5B maintained fixed memory usage and outperformed ICL and RAG, though the evidence is currently limited to three synthetic retrieval tasks.

2026-08-19

最新模型

TokEval Screens Tokenizers Using 18 Intrinsic Metrics, With Information-Theoretic Metrics Correlating With Language Modeling Performance by Up to 0.80

TokEval incorporates numeric place value, program AST boundaries, UTF-8 integrity, and cross-lingual cost into tokenizer evaluation. Controlled experiments show that some metrics can predict specific downstream capabilities, although the results are currently limited to a single 1.27B-parameter architecture.

2026-08-19

代理工具

Agent Lightning v1.0 Brings Real-World Agent Frameworks Directly Into Reinforcement Learning, Raising Qwen3.5-9B to 56.4% on SWE-bench Verified

Microsoft rewrote Agent Lightning to capture complete interactions from existing agent frameworks through a model API proxy layer, eliminating the need to move tool loops into the training engine. Using 6,000 training examples, the team reports that Qwen3.5-9B improved from 41.8% to 56.4% on SWE-bench Verified.

2026-08-19

推論系統

TileMix Mixes FP16 and INT8 Within the Attention Matrix, Reaching 2.22× FlashAttention Throughput for 4K Prefill on A100

Rather than removing token connections, TileMix selects either the FP16 or INT8 Tensor Core path for each attention-score tile within a single fused kernel. It reaches 31.80K token/s on a 4K LLaMA 3.2 3B benchmark, though the results are currently limited to A100, static routing, and the prefill stage.

2026-08-19

代理工具

AeroCopilotBench Turns Flight Procedures Into Executable State Machines, Yet the Best Agent’s Safe Pass Rate Is Still Only 72.6%

The new benchmark uses 73 abnormal cockpit tasks to test whether agents can observe state, operate systems, and comply with hard safety constraints throughout—not merely answer aviation knowledge questions. GPT-5.6 Sol achieved the highest success rate, but similar knowledge scores did not correspond to similar task success rates, showing that static question answering cannot replace interactive safety testing.

2026-08-19

代理工具

Reconstruction Hides Papers and Provides Only Older References; Multi-Model Review Raises Research-Idea Match Rate from 14.6% to 35.6%

A new benchmark asks models to reconstruct a hidden paper’s core research idea using only references that existed before its publication. Cross-review by four models and Swiss-system selection improve the match rate, but the candidate count, judge pool, and compute budget are not aligned with the single-model setting.

2026-08-19

最新模型

FabriMAE 以視覺注意力熵判斷機器人動作可靠度，十分支選擇令成功率升至 86.8%

FabriMAE 從 VLA 模型內部的視覺注意力熵產生可靠度分數，不需另訓練驗證器或重複執行機器人軌跡。PI0.5 在 LIBERO-Plus 的成功率增加 1.1 個百分點，但目前證據仍限於模擬環境，論文連結的程式庫亦尚未公開存取。

2026-08-18

代理工具

PDDLCoder Lets Agents Iteratively Fix PDDL, Achieving an 89.6% Executable Success Rate on 106 Unseen Planning Problems

Rather than requiring a language model to directly generate long-horizon action sequences, PDDLCoder first constructs a PDDL world model that can be verified by a symbolic planner. DeepSeek V4 Flash solved 95 of 106 problems across four held-out domains, but results for different base models ranged from 89.6% to zero, showing that the framework’s effectiveness depends heavily on a model’s ability to follow formatting requirements and use tool feedback.

2026-08-18

代理工具

Kozuchi Takes the Same 27B Code-Repair Agent to Java, Solving 41 of 128 Multi-SWE-bench Tasks

Fujitsu ported its fine-tuning-free Qwen3.5-27B setup, eight-stage repair workflow, and cross-agent test selector to Java unchanged. It solved 374 of 500 tasks on the Python SWE-bench Verified benchmark, but achieved only a 32.03% success rate on Java, showing that agent-framework portability does not imply equal language proficiency.

2026-08-18

最新模型

PertMind Trains a 4B Model on Cellular Perturbation Endpoints, Turning 100 Million Single-Cell Profiles into Reinforcement Learning Rewards

Starting from Qwen3-4B, PertMind combines gene responses, pathway direction, and output formatting into computable rewards, eliminating the need to manually author large numbers of biological reasoning traces. The model transfers zero-shot to inverse perturbation identification and dual-perturbation tasks, but its natural-language mechanistic explanations still cannot be treated as causal proof.

2026-08-18

推論系統

R³-Bench Gives Six Problems a Shared Reasoning Budget; Models Fall Below Their Own Offline Upper Bound in 71 of 72 Test Settings

The new evaluation no longer provides a separate allowance for each problem. Instead, models must allocate tokens or tool calls across six problems. Even though all six models can solve the problems individually, under a shared budget they generally fail to direct compute toward the problems with the highest expected payoff.

2026-08-18

代理工具

RelArena-α Uniformly Reruns 21 Relational Data Tasks, With TabPFN-Rel Leading Among Models Using Standardized Tuning

Prior Labs has open-sourced a unified workflow for data splitting, hyperparameter tuning, and evaluation, enabling a fresh comparison of graph models, relational Transformers, and tabular methods. TabPFN-Rel shows that aggregating relational data before passing it to a tabular foundation model remains competitive, although the CPU cost of feature synthesis is not fully included in the leaderboard’s timing results.

2026-08-18