洞察 AI 业界
最新进展

智能指数 - · 输出吞吐 74.92 tok/s

智能指数 - · 输出吞吐 65.48 tok/s

智能指数 41.7 · 输出吞吐 - tok/s

智能指数 - · 输出吞吐 78.02 tok/s

智能指数 - · 输出吞吐 70.02 tok/s

智能指数 53.4 · 输出吞吐 73.69 tok/s

智能指数 28.9 · 输出吞吐 - tok/s

智能指数 51.1 · 输出吞吐 206.84 tok/s

智能指数 41.9 · 输出吞吐 51.17 tok/s

智能指数 13.5 · 输出吞吐 - tok/s

智能指数 59.9 · 输出吞吐 70.41 tok/s

智能指数 20.6 · 输出吞吐 89.02 tok/s

# Release v5.13.0 ## New Model additions ### KimiK 2.5, 2.6, and 2.7 <img width="1097" height="400" alt="image" src="https://github.com/user-attachments/assets/c24d2232-a9b4-413b-a2c8-58d013b6dfbd" /> This release includes the architecture for Kimi 2.5 which is used by 2.5-2.7: Kimi K2.5 is an open-source, native multimodal agentic model that advances practical capabilities in long-horizon coding, coding-driven design, proactive autonomous execution, and swarm-based task orchest

# vLLM v0.24.0 Release Notes ## Highlights This release features 571 commits from 256 contributors (77 new)! * **MiniMax-M3**: Added support for the new **MiniMax-M3** model (#45381), with a fast follow-on of BF16/FP8 indexer via MSA (#45892), MXFP4 support (#45896), FP8 sparse GQA (#45744), and extensive AMD/ROCm tuning — mxfp8 MoE/linear on gfx950 (#45725), fp8_per_channel for bf16 weights on MI300X (#45854), FP8 KV-cache fix (#45720), and packed-modules mapping (#45794). A MiniMax-M2

# Highlights New Model Support: [GLM-5.2](https://docs.sglang.io/cookbook/autoregressive/GLM/GLM-5.2), [LiquidAI LFM2.5](https://docs.sglang.io/cookbook/autoregressive/LiquidAI/LFM2.5), [Kimi-K2.7-Code](https://docs.sglang.io/cookbook/autoregressive/Moonshotai/Kimi-K2.7-Code), [Poolside Laguna-M.1](https://docs.sglang.io/cookbook/autoregressive/Poolside/Laguna-M.1), [DiffusionGemma](https://docs.sglang.io/cookbook/autoregressive/Google/DiffusionGemma), Zyphra ZAYA1, MiMo-V2-ASR **DeepSeek-

# Patch release v5.12.1 Updated the lower bound for PEFT and a fix for auto tokenizer to properly resolve the mistral tokenizer (when `mistral-common` is installed). This is similar to v.5.10.3 minus the fixes that were already included in the main release - vLLM will first target 5.10.3 :hugs: * Fix `peft` lower bound #46605 by @hmellor (#46605) * mistral common backend fix #46667 by @itazap (#46667) **Full Changelog**: https://github.com/huggingface/transformers/compare/v5.12.0...v5

# Patch release v5.10.4 Update: Note that on pypi `5.10.3` doesn't exist and this this saved under `5.10.4` (so essentially a minor version skipped). Sorry about that, that's on me. Just wanted to clarify to make this less confusing! A few fixes needed for vLLM to sync with transformers :hugs: * [fix] regression introduced by #45534 #46456 by @eustlb (#46456) * Fix {image/video/audio}_token_ids in ProcessorMixin #46500 by @hmellor (#46500) * Fix InternVL models #46524 by @hmellor (#465

# vLLM v0.23.0 Release Notes Please note that Minimax M3 is not yet supported in this version. Please follow [vLLM recipe](https://recipes.vllm.ai/MiniMaxAI/MiniMax-M3) for usage guides for M3. ## Highlights This release features 408 commits from 200 contributors (63 new)! * **DeepSeek-V4 matures across backends**: Following its introduction in v0.22.0, DeepSeek-V4 received another large hardening and optimization pass. Its sparse MLA metadata is now decoupled from DeepSeek-V3.2 (#44

## Highlights **New Model Support**: - **Autoregressive**: [Nemotron 3 Ultra](https://docs.sglang.io/cookbook/autoregressive/NVIDIA/Nemotron3-Ultra) (Day-0, [blog](https://www.lmsys.org/blog/2026-06-04-nvidia-run-nemotron-3-ultra/)), [Step-3.7-Flash](https://docs.sglang.io/cookbook/autoregressive/StepFun/Step-3.7-Flash), Command A+ - **Diffusion**: [Cosmos3](https://docs.sglang.io/cookbook/diffusion/Cosmos/Cosmos3), [LingBot-World](https://docs.sglang.io/cookbook/diffusion/LingBot-World/L

# Release v5.12.0 ## New Model additions ### MiniMax-M3-VL <img width="886" height="583" alt="image" src="https://github.com/user-attachments/assets/ae9dd96f-6877-4531-a06b-a756686f24e5" /> MiniMax-M3-VL is the vision-language member of the MiniMax-M3 family that pairs a CLIP-style vision tower with 3D rotary position embeddings with the MiniMax-M3 text backbone. It uses a mixed dense/sparse Mixture-of-Experts decoder with SwiGLU-OAI gated experts and a lightning indexer for block-

# Release v5.11.0 ## New Model additions ### DiffusionGemma <img width="1240" height="700" alt="image" src="https://github.com/user-attachments/assets/5081e449-6374-4076-bd96-d295c8334ca4" /> DiffusionGemma is engineered to reduce the sequential bottlenecks of standard causal language models by employing an encoder-decoder architecture specifically optimized for inference speed. During inference, DiffusionGemma leverages multi-canvas sampling, where rather than generating one token

## Highlights This release features 8 commits from 6 contributors (1 new)! v0.22.1 is a patch release on top of v0.22.0 with targeted bug fixes plus a couple of additions: new model support for JetBrains' Mellum v2, zentorch-accelerated quantized linear inference on AMD Zen CPUs, and fixes for multi-node Ray data-parallel serving, DeepSeek-V4 initialization, and a few model-loading regressions. ### Model Support * New model: JetBrains' **Mellum v2**, an open-weights Mixture-of-Experts

# Patch release v5.10.2 There was a big bug in the model conversion of models related to clip, this affected models like sam3 and others. Please make sure to update :pray: * Fix conversion for clip models by @zucchini-nlp (#46406) **Full Changelog**: https://github.com/huggingface/transformers/compare/v5.10.1...v5.10.2

# Release v5.10.1 v5.10.0 was yanked as we publish on a corrupted branch. Sorry everyone, this happens when we rush a release!!! ## New Model additions ### Gemma4 unified+ Gemma4 MTP <img width="2000" height="400" alt="image" src="https://github.com/user-attachments/assets/5e3ee940-f78d-4343-ac7a-889930800aa6" /> Gemma 4 12B Unified is an **encoder-free** multimodal model with pretrained and instruction-tuned variants. Unlike [standard Gemma 4](./gemma4), which uses dedicated encoder

LLMs memorize sensitive training data, including personally identifiable information (PII), creating a pressing need for reliable post hoc removal methods. Unlearning has emerged as a promising solution, with state-of-the-art(SOTA) methods often following a localize-first, unlearn-second paradigm that targets specific model parameters. However, existing benchmarks evaluate unlearning solely at the output level, leaving open the question of whether unlearning truly erases knowledge from a model's

Many everyday programming tasks resist clean rule-based implementation, such as alerting on important log lines, repairing malformed JSON, or ranking search results by intent, and are increasingly outsourced to large language model APIs at the cost of locality, reproducibility, and price. We propose fuzzy-function programming: compiling such a function from a natural-language specification into a compact, locally-executable neural artifact. We instantiate this paradigm with Program-as-Weights (P

Despite alignment training, LLMs remain prone to generating unsafe outputs at deployment time. Monitoring outputs online and raising an alarm when safety can no longer be assumed is therefore critical. We study a simple real-time monitor that turns a verifier signal from an external model into an alarm decision by thresholding, with the threshold calibrated via risk control. In experiments on mathematical reasoning and red teaming datasets, we show that this simple design is competitive with mor

LLM agents will increasingly act in socially structured settings where role, audience, and relational context can shape what is advantageous or costly to say. We study whether such social structure, without any explicit objective in the prompt, changes what an agent expresses publicly relative to an off-the-record (OTR) channel elicited under the same condition. We introduce a dual-channel debate framework in which agents produce public utterances that enter the shared history alongside OTR resp

Long-form TV dramas present a formidable challenge for comprehensive video understanding, where deciphering complex storyline often relies on \textbf{speaker recognition}, the task of accurately attributing each spoken utterance to its respective character. In this paper, we advance this field through two primary contributions. (1) We introduce \textbf{DramaSR-532K}, a large-scale benchmark comprising 532K annotated dialogue lines across more than 900 unique characters, necessitating the integra

Models trained via Contrastive Language-Image Pretraining (CLIP) serve as the foundational vision encoders for most modern Large Vision Language Models (LVLMs). Despite their widespread adoption, CLIP models exhibit a critical yet underexplored failure mode: irrelevant text appearing within images confounds visual representations, biasing them toward lexical meaning rather than true visual semantics. This robustness issue, commonly described as a Typographic Attack (TA), exposes a vulnerability

Large vision-language models can reason over multimodal inputs by generating textual chains of thought (CoT). A key capability exhibited in CoT reasoning is self-reflection: revisiting earlier decisions and correcting previous errors. However, existing LVLMs often fail to properly attend to visual inputs during reflection, limiting their ability to translate feedback into grounded corrections, especially for out-of-distribution images. To address this issue, we propose a novel reinforcement lear

Narration is central to the audiobook listening experience, shaping how listeners engage with and understand the content. This work explores how narration qualities shape an audiobook's appeal, noting that their effects can vary by genre, title, and audience. We extract vocal and acoustic features (e.g., tone, pace, loudness) from LibriVox using pre-trained audio models and analyse their relationship with consumption data (specifically, view-rate) and their interplay with genre and title. Despit

Software tests and code evolve together: a code change should be followed by new or updated tests that record the new software behavior. Yet existing test generation and update benchmarks often isolate the test from the code change, and rely on static metadata that does not verify whether a test is executable or semantically tied to the code change. This makes it difficult to evaluate whether a test automation agent understands how a code change should propagate into the test suite. We introduce

Large Language Model (LLM) social simulations are a promising research method, but they are not yet faithful enough to be adopted widely. In this work, we investigate whether the current scaling paradigm in language modeling is likely to close these gaps, or whether simulation fidelity is orthogonal to general capabilities and therefore deserving of more research attention. We use scaling laws to study the relationship between LLMs' compute scale, general capability benchmarks, and the fidelity

Language models are increasingly used to quantify cultural phenomena, but what makes such measurement distinctively cultural? This paper argues that NLP work on culture is a material-discursive practice: the apparatus -- model, data, annotation, evaluation -- participates in constituting the cultural reality it measures, rather than passively recording it. Drawing on Karen Barad's concept of the agential cut -- the contingent boundary between phenomenon and instrument -- I show that the apparatu

Autonomous agents are increasingly expected to improve executable policies through feedback, yet existing evaluations often collapse this process into a final score or confound it with open-ended software-engineering progress. We introduce Autonomous Policy Evolution, a controlled evaluation setting in which a harness-model agent repeatedly edits an executable policy system under a fixed interaction budget. We instantiate this setting in EvoPolicyGym, a benchmark built from compact interactive R