Latest news, beside the leaderboard

On Oct 18 Anthropic announced the Anthropic Public Benefit Fund, a wholly owned LLC. Subject to stockholder approval, 5% of shares are committed now, and the board may add up to 2% for every 50% rise in company value above a set baseline. With the co-founders' pledge to donate 80% of their equity and an employee equity-donation match that about half of staff use, roughly 20% of Anthropic's shares go to the public good. The fund's equity will not be used for lobbying or political giving; it may fund large nonprofit donations or direct distributions, for example to address AI's labor-market impact.

vLLM shipped v0.32.0 on Oct 18 with 814 commits from 363 contributors (121 new). Highlights: a default-served POST /v1/systemone structured-decision endpoint and a generic hf tool/reasoning parser; NVFP4 KV cache on SM8x/SM90/SM12x and FP8 KV by default for GLM-5.x on SM100; Model Runner V2 gains a GPU n-gram speculator and Elastic EP and becomes the ROCm default; three security advisories are fixed, default CORS is restricted, and Python 3.10 is dropped.

Addressing the representation gap between high-level VLM semantics, video world models, and low-level control in Vision-Language-Action (VLA) architectures, PINT (arXiv:2610.16355) introduces learnable chunk-level Predictive Intent Tokens to reach 15.5% on RoboTwin 2.0-Hard without embodied pretraining (4.8x over StarVLA's 3.2%). Alongside PAIR (arXiv:2610.09016) and PKU's SimForcing (arXiv:2610.06598), these frameworks bridge perception and action with zero inference overhead.

Three new architectures advance 4D perception, 3D PBR texturing, and real-time autoregressive video diffusion: Gen4P (arXiv:2610.22371) repurposes a video generator into a single-pass feed-forward 4D perceiver without iterative denoising; HUST and Huawei's SUPRA-TEX (arXiv:2610.22352) reaches 32.08 dB Shaded PBR PSNR on DTC-Tex; and HKUST's ForcingBooster (arXiv:2610.12625) cuts patch denoising NFE by 50.19% for a 1.30x end-to-end speedup across Wan2.1 14B and LongLive.

Alibaba Taobao & Tmall Group detailed CommerceGen-Studio (arXiv:2610.20603, 52 HF upvotes), a production multi-agent generation system serving millions of images daily within seconds. Its 6B SKU-DiT trained via Reinforce-Distill-Reinforce (RDR) scores 9.43 (single-image) and 9.45 (multi-image) on CommerceGen-Bench, outperforming 20B open models and closed-source Nano Banana Pro. Meanwhile, Microsoft Research introduced Mine, Sample, Render (MSR, arXiv:2610.21185) for training-free test-time diversity alignment.

Researchers from Tsinghua University open-sourced Dynamic Band Splitting (DBS, arXiv:2610.14993, 134 upvotes on Hugging Face Papers), replacing fixed predefined frequency partitions in band-split audio source separation models with content-conditioned Top-K boundary selection and adaptive pooling under a fixed token budget. Integrating DBS into BSRNN yields a +29.0% relative improvement in average SI-SDR on the DnR v2 benchmark.

Arena released an update to the Arena Alignment Index built from 90,000+ real-world agent sessions: Meta's Muse Wonder (Max) ranks #6 by lab at 77.5% (2.09% unauthorized action rate, 14.41% deceptive completion rate), Reflection AI's open-weight Beam-1.0-501B-A23B debuts at #7 with 75.3% (4.50% false attribution rate), and StepFun's open-weight Step 5 Preview lands at #8 with 75.1% (2.11% unauthorized action rate).

Two new 3D vision and generative compression frameworks solve long-standing efficiency and quality bottlenecks: Zhejiang University's InfiniSplat v2 (arXiv:2610.20341, 118 upvotes on HF) decouples geometry and Gaussian attribute prediction with a frozen geometry branch and large-angle supervision, slashing mean LPIPS by 25.8% and DISTS by 32.9% across six 3D datasets. Meanwhile, Technion's H-DDCM (arXiv:2610.18125) replaces dense Gaussian codebooks in training-free diffusion compression with implicit Walsh-Hadamard dictionaries—requiring O(d log d) compute and O(d) memory—to enable near-real-time streaming video compression on a single GPU.

Three new studies tackle verification, evaluation, and RL alignment for rule-governed generative world models: OmniVerifier-C1 (arXiv:2610.22355, 32 upvotes) uses Decoupled RL to train a decomposed graded visual verifier that reaches 0.684 on ViVerBench and boosts FLUX.1-dev GenEval from 0.68 to 0.79 while enabling Maze and Sudoku world modeling; Meituan's WorldCast benchmark (arXiv:2610.20731, 30 upvotes) evaluates 12 models across 320 multi-step world state prediction cases, finding GPT-Image-2.5 System at 72.8 vs. 42.8 for Nano Banana Pro and 20.1 for the top open model; and NUS's Routing Policy Optimization (RPO, arXiv:2610.22282, 23 upvotes) turns attention routing into an internal exploration action to train video/image diffusion RL 3.1x–4.2x faster than Flow-GRPO and DanceGRPO.

Three new evaluation and interpretability studies resolve critical measurement blind spots: POSTECH and KAIST researchers (arXiv:2610.20993) prove across 41 checkpoints and 12 benchmarks that multiple-choice pooled selection-rate SD equals the exact product of Flip Rate and Positional Skew, revealing that forward-reverse order aggregation lowers flip rates partly by locking onto wrong answers; semCER (arXiv:2610.20461) open-sources 61 million LLM-judged grapheme edits across 16 languages for severity-weighted OCR evaluation; and IISc/UVA researchers (arXiv:2610.20680) show that while SAE reconstruction converges with little data, strict 1-to-1 feature reproducibility on Gemma-2-2B BatchTopK jumps 14.5x from 2% at 5M tokens to 29% at 50M tokens.

A vLLM blog post by NVIDIA engineer Benjamin Chislett details an agent-assisted CuTe DSL decode path for Kimi K3 on NVIDIA GB300 that fuses attention, MoE and NVLink communication, cutting per-layer time from 130 µs across 23 kernels to 69 µs across 13. With DSpark speculative decoding on 16 GPUs it reaches 842 tok/s at batch 1, 1.73x stock vLLM; PR #61867 was merged on Oct 16.

vLLM and NVIDIA unveiled a unified CuTe DSL decode megakernel (~12,000 LoC authored by NVIDIA CentML's Benjamin Chislett) for Moonshot AI's frontier Kimi K3 model on NVIDIA GB300 NVL72. By fusing attention, MoE, and cross-GPU communication, the optimization cuts per-layer latency from 130 µs (23 kernels) to 69 µs (13 kernels), delivering 842 tokens/s at batch=1 with DSpark speculative decoding (1.73x faster than stock vLLM) and up to 1.8x higher P90 per-user interactivity (403.2–423 tok/s/user) on SemiAnalysis AgentX, with DeepSeek-V4.1-Flash (PR #61868) and GLM-5.3 support coming next.

Three new frameworks advance web and multimodal agent post-training: Salesforce AI Research's CLIFT (arXiv:2610.06829) certifies URL-conditional self-verification questions for dense RL rewards and judge-free Conformal Trajectory Selection (CTS), propelling Gemma-4 31B to a 74.6% SOTA on WebArena Infinity (beating Gemini 3 Flash at 70.1%) and lifting GPT-5.5 to 53.7% on VisualWebArena. Meanwhile, MMSearch-OPD (arXiv:2610.22240) boosts 9B multimodal deep search by +15.3 points to 60.2%, and Northeastern University's SCOPED-8B (arXiv:2610.22217) uses skill-decoupled on-policy expert distillation to reach 73.13% across 8 benchmarks—surpassing Qwen3-VL-32B (72.01%) and Claude Opus 4.6 (72.86%).

Four new studies advance interactive video world models and streaming video agents: NUS's training-free Velox router (arXiv:2610.22311) uses 3D camera frustum-slab correspondences on H200 GPUs to accelerate chunk-48 attention by 4.36x and denoising by 2.43x with negligible quality loss; Nankai University's DeltaStreamer (arXiv:2610.19849, 39 upvotes) combines change-aware memory with targeted frame replay to reach 66.07% on OVO-Bench; HKPolyU's CoWorld Reasoner (arXiv:2610.21910) co-evolves a VLM-VWM tool harness for +13.0% relative gains on ERQA; and Zhejiang University releases the 260-case WorldMemBench (arXiv:2610.19063).

Two mechanistic interpretability studies uncover how LLMs express and localize internal knowledge: University of Cambridge researchers (arXiv:2610.21025) introduce parameter-trajectory crosscoding on checkpoint weights, discovering that hidden states encode task knowledge before the unembedding readout matrix learns to express it (reorganizing at 0.4-1.1% of Pythia pretraining), and that raising the readout learning rate closes this gap earlier. Meanwhile, University of Copenhagen researchers (arXiv:2610.22127) apply paired-activation crosscoders to three open-weight Mixture-of-Experts (MoE) LLMs, isolating monosemantic personalization latents and proving via causal interventions that boosting specific routed experts recovers personalized model behavior.

Ollama launched US- and EU-hosted Ollama Cloud with strict zero data retention and zero training, adding immediate support for DeepSeek-V4.1-Flash-1017 via one-line launch commands for Claude Code and OpenCode alongside Ollama v0.40.6. Meanwhile, SemiAnalysis revealed that GLM-5.3 was built entirely by post-training the previous GLM-5.2 base model without pretraining from scratch, achieving a 'night and day difference' in cybersecurity and agentic capabilities.

Tackling step-level verification in mathematical reasoning and in-context copying bottlenecks in Transformers, NYU/UCLA/Stanford researchers introduce LeanPRM (arXiv:2610.17697)—using the Lean 4 theorem prover to automatically label 152K steps across 10,634 natural-language proofs with zero human annotation, boosting Qwen3-4B step accuracy on LeanPRM-Bench from 48.2% to 74.7% (vs. 60.3% for PRM800K). Meanwhile, Cornell University's Copy-Paste Language Models (CPLMs, arXiv:2610.19247) add a lightweight Copy-Paste head (~3% compute overhead) that relieves the LM head gradient bottleneck, boosting LAMBADA accuracy by +2.5-4.0 points and allowing a w=128 sliding-window CPLM to match full-attention perplexity.

Three new open-source embodied AI frameworks overcome observation-action overfitting in Vision-Language-Action (VLA) policies: Zhejiang University's 2B VLA-Δ (arXiv:2610.21306, 54 upvotes) supervises explicit physical state transitions with zero inference overhead to set a 98.4% SOTA on LIBERO; SJTU and HKU's OneWM (arXiv:2610.21095, 40 upvotes) unifies causal world and action modeling to reach 36.68% on RoboCasa365 (over 2x π0.5 at 16.90%); and UIUC/Amazon's CoDA (arXiv:2610.22362) co-evolves simulation data engines and action policies via Text2potential shaping, raising LIBERO-Pro success from 25% to 84% under position perturbations and from 1% to 76% under instruction perturbations.

SemiAnalysis published the architectural breakdown behind the 1.3B tokens per dollar TCO record on AgentX: vLLM and Inferact utilized Prefill/Decode disaggregated serving at hyperscaler volume pricing to achieve 4x better cost-efficiency than an aggregated NVIDIA B300 configuration. Concurrently, open-source coding agent Cline reported that DeepSeek-V4.1-Flash-1017 beats GPT-5.6-Sol on Terminal-Bench at ~17x lower cost (+35% over prior V4.1-Flash) purely by scaling RL training across multiple coding agent harnesses.

On 2026-10-17 OpenCode disclosed an overnight breach tied to a large-scale distillation attack affecting about 1.4% of Console users. Affected accounts were emailed; no email means not impacted; only users with external providers linked. Remediation is ongoing with Anthropic, OpenAI and Google. No official blog or GitHub advisory yet; the status page listed no dedicated security incident.