ZeroHour
DashboardNewsCVEsExploitsIncidentsAIBriefAnalyticsEntitiesWatchlistIOCsAsk the feedSourcesFeeds & APIKiosk mode

Search: “StyleSmuggler implant”

31 stories

Searchkeywordsmart
DomainAllCyberAIWindow24h3d7d30dAllIn the wildImportant onlyShow duplicates
CategoryAllExploit / PoC 7
SeverityAllcritical 7high 0medium 0low 0info 0

Smart search ranks by meaning as well as keywords (one row per story, last 45 days).

StyleSmuggler: Magento and Adobe Commerce 0-day RCE (CVE-2026-75650) under active attack

Sansec details actively exploited StyleSmuggler 0-day (CVE-2026-75650, CVSS 10.0) unauthenticated RCE in Magento and Adobe Commerce, patched by Adobe hotfix APSB26-146.

Sansec is investigating StyleSmuggler, an actively exploited unauthenticated remote code execution chain in Magento Open Source and Adobe Commerce, now tracked as CVE-2026-75650 with CVSS 10.0. Adobe released hotfix VULN-39341 via APSB26-146 (priority 1) on September 7 for versions 2.4.4 through 2.4.9, but stores were being exploited for roughly three days before the fix existed. The implant is a Rust backdoor that disguises itself as kworker, fc-cache, or chronyd processes and exfiltrates host data in MessagePack records sent as fake NTP replies over UDP port 123. Adobe advises rotating the encryption key and every credential it protected, and Sansec stresses patching does not clean already-compromised stores.

Sansec (Magento / e-commerce security) · 12d agoExploit / PoC in the wildCVE-2026-756501

StyleSmuggler: The Magento Zero-Day Behind New Store Attacks

Sansec reports actively exploited Magento/Adobe Commerce zero-day StyleSmuggler enabling unauthenticated RCE and Rust backdoor installation on fully patched stores since September 4.

Sansec discovered StyleSmuggler, an unpatched zero-day in Magento Open Source and Adobe Commerce, affecting all current versions including 2.4.7, 2.4.8 and 2.4.9, with attacks observed since September 4. The two-stage attack poisons Magento's template system via the styles property and executes the injected PHP during 'Payment Transaction Failed Reminder' email rendering, working even when email delivery fails and when sessions are moved to Redis. Successful compromise installs a lightweight Rust backdoor disguised as fc-cache or chronyd that beacons every 60 seconds with 48-byte UDP packets to NTP port 123 at ntp.timesync.to. A second attacker deployed a PHP web shell in product-image cache directories, hidden behind 404 responses unless a correct X-Cache-Token header is present.

Security Affairs · 9d agoExploit / PoC in the wild

Unpatched Magento and Adobe Commerce Zero-Day Exploited to Backdoor Online Stores

Sansec reports active exploitation of an unpatched zero-day, dubbed StyleSmuggler, in Magento and Adobe Commerce, letting attackers backdoor online store servers without authentication.

Sansec disclosed on September 5 that attackers are exploiting an unpatched vulnerability, named StyleSmuggler, in Magento Open Source and Adobe Commerce to achieve unauthenticated remote code execution, with attacks observed starting September 4. All current versions including 2.4.9 are affected, and Adobe has published no advisory, CVE, patch, or workaround, with the next scheduled security release on September 8. Hosting firm Disrex Group independently confirmed two compromised stores (running 2.4.8 and 2.4.7-p2), both breached within the roughly eight-hour window before Sansec's blocking rules went live. The implant is a ~1.9 MB statically linked Rust binary disguised as a Linux kworker process, re-adding a cron entry every five minutes and in one case reading Magento sessions directly from Redis with no outbound traffic.

The Hacker News · 11d agoExploit / PoC in the wild

StyleSmuggler (CVE-2026-75650): Frequently asked questions about Adobe Commerce and Magento zero-day

Critical unauthenticated RCE zero-day CVE-2026-75650 (StyleSmuggler) in Adobe Commerce and Magento is actively exploited; Adobe shipped hotfix VULN-39341 on September 7.

CVE-2026-75650 is a CVSS 10.0 unauthenticated remote code execution flaw in Adobe Commerce 2.4.4-2.4.9, Adobe Commerce B2B 1.3.3-1.5.3, and Magento Open Source 2.4.6-2.4.9, exploited via malicious style properties that inject PHP code executed through a transactional email template. Active exploitation began September 4, 2026, three days before Adobe released Hotfix VULN-39341 (APSB26-146) on September 7, with multiple victim stores confirmed by Sansec and Disrex across at least two distinct campaigns. Attackers deploy a persistent implant at ~/.local/share/.gvfsd/gvfsd-user masquerading as kworker, fc-cache, or chronyd, sustained by a cron job, while a second operator dropped a PHP web shell in product image caches. Tenable classified it as a Vulnerability of Interest; it is not yet in CISA KEV as of September 8, and Adobe also recommends rotating encryption keys and all protected credentials.

Tenable Blogupdated · 8d agofirst · 8d agoExploit / PoC in the wild 2 sourcesCVE-2026-75650CVE-2025-54236CVE-2024-34102+1 CVEs

Adobe Commerce Zero-Day Exploited to Backdoor Online Stores

Sansec found attackers exploiting an Adobe Commerce/Magento zero-day (StyleSmuggler) since September 4 to deploy Rust backdoors on online stores.

Threat actors are actively exploiting a zero-day RCE in Adobe Commerce and Magento 2.4.7-2.4.9, injecting PHP code via generated failure reports and executing it through Magento's payment-failure email, with no user interaction required. Exploitation began September 4 and succeeds even against stores running the July and August 2026 patches; the Rust backdoor disguises itself as kworker/u:8:0 or fc-cache and hides C&C communication inside fake NTP replies. Adobe's September 8 Patch Tuesday updates may not include a StyleSmuggler fix.

SecurityWeek · 9d agoExploit / PoC in the wild

CVE-2026-75650: StyleSmuggler — Critical RCE in Adobe Commerce and Magento

Akamai research details CVE-2026-75650 (StyleSmuggler), a critical remote code execution flaw in Adobe Commerce and Magento.

Akamai Security Research published analysis of CVE-2026-75650, dubbed StyleSmuggler, a critical remote code execution vulnerability affecting Adobe Commerce and Magento platforms. The research appears in Akamai's security blog. Full technical details were not available from the item metadata.

Akamai Blogupdated · 14h agofirst · 2d agoVulnerability 11 sourcesCVE-2026-75650

StyleSmuggler: Unpatched Magento and Adobe Commerce Zero-Day Exploited

Attackers are actively exploiting an unpatched zero-day dubbed StyleSmuggler in Magento Open Source and Adobe Commerce for unauthenticated remote code execution.

A zero-day vulnerability tracked as StyleSmuggler affects Magento Open Source and Adobe Commerce and permits unauthenticated remote code execution. The flaw is unpatched and exploitation is ongoing in the wild. E-commerce stores running these platforms are at direct risk of compromise until a fix is released.

SOCRadar · 9d agoExploit / PoC in the wild

Adobe Commerce max-severity bug comes under active attack

Attackers are actively exploiting CVE-2026-75650 (CVSS 10.0), an unauthenticated RCE zero-day in Adobe Commerce/Magento, deploying a Rust backdoor on e-commerce servers.

Sansec identified active exploitation of a CVSS 10.0 zero-day (CVE-2026-75650) in Adobe Commerce and Magento Open Source versions 2.4.4 through 2.4.9, dubbed StyleSmuggler. The attack abuses crafted style properties in Magento template processing to execute PHP while rendering a 'Payment Transaction Failed Reminder' email, then launches a Rust backdoor connecting to C2 at 99.84.67.186 disguised as NTP traffic. Exploitation began September 4, 2026; Adobe released emergency hotfix VULN-393411, but Sansec also observed a second attacker deploying a PHP web shell in the product-image cache.

CSO Online · 8d agoExploit / PoC in the wildCVE-2026-75650

Weekly Cybersecurity Newsletter Bulletin – Microsoft 0-day, FortiOS, PAN-OS Flaw, Revolut Data Breach, and 20+ Stories

Weekly roundup: Microsoft patches 973 flaws including two actively exploited zero-days; FortiOS CAPWAP flaw deploys PivotC2 RAT; PAN-OS root RCE disclosed.

Microsoft's September 2026 Patch Tuesday fixed 973 vulnerabilities, including two zero-days under active exploitation: CVE-2026-85880 (Windows ALPC) and CVE-2026-81963 (Windows Update Stack), both elevation-of-privilege bugs. SOCRadar reported active exploitation of CVE-2025-25249 (CVSS 9.8) in FortiOS CAPWAP, deploying a Node.js RAT called PivotC2 that exfiltrates Exchange mailboxes to Wasabi cloud storage; 178 devices were compromised out of 30,000 scanned IPs, attributed to a Russian-speaking financially motivated group. Palo Alto disclosed CVE-2026-0310, a 9.2-rated buffer overflow enabling root code execution on PA-Series firewalls, and Fortinet disclosed CVE-2026-84393, a ZTNA certificate validation MITM flaw. Cyera also revealed CVE-2026-6471 ('PostGREShell'), a 12-year-old PostgreSQL logical-decoding flaw allowing code execution via REPLICATION-privileged accounts.

Cyber Security News · 3d agoExploit / PoC in the wildCVE-2026-85880CVE-2026-81963CVE-2025-25249+3 CVEs3

Magento StyleSmuggler zero-day exploited to deploy Linux backdoor

StyleSmuggler zero-day in Magento and Adobe Commerce is actively exploited in the wild to deploy a Rust-based Linux backdoor; fix not yet released.

A zero-day dubbed StyleSmuggler affecting all versions of Magento and Adobe Commerce is being actively exploited, with the first observed incident on September 4 on a fully patched site. The exploit injects PHP code via Magento's template system through a fake 'failed-payment' email to achieve code execution, installing a Rust-based backdoor disguised as kworker/u:8:0 or fc-cache with a 30-minute cron job for persistence. The backdoor communicates via TLS/WebSockets or UDP port 123 masquerading as NTP traffic, checks TracerPid to detect tracing, and can receive remote commands. Adobe confirmed it is working on a fix with no timeline; Sansec recommends disabling GraphQL as mitigation and monitoring for suspicious cron entries and kworker or fc-cache processes. Magento is installed on more than 160,000 websites, including 14,000 of the top 1 million sites.

BleepingComputer · 9d agoExploit / PoC in the wild

Viggle/Viggle-Animate — new model trending #28 on Hugging Face

Viggle released Viggle-Animate, a 33.1B MiniMax-H3 finetune replacing video characters from one repainted frame, rendering 124 frames in 26 seconds on one GPU.

Viggle-Animate replaces the character in a video using only a driving video and one of its own repainted frames, with no pose estimator, segmentation mask, face tracker, or text encoder. It is a 33.1B full finetune of MiniMax-H3's ref2va transformer, jointly distilled with DMD across two teachers split by noise level, so rendering takes three forward passes per clip. On a B200 GPU it renders 124 frames in 26 seconds, 6.1x faster per clip than Wan2.2-Animate-14B in matched comparisons. The method assumes no person-specific representation, so it generalizes beyond humans; a demo, research write-up, and ComfyUI nodes are available.

Hugging Face trending models · 16d agoModel release1

Comfy-Org/YuE2 — new model trending #30 on Hugging Face

m-a-p's YuE2-3B music generation model and SheetSage2 audio encoder are repackaged in bf16 for ComfyUI and trending #30 on Hugging Face.

Comfy-Org published repackaged bf16 safetensors files for m-a-p's YuE2-3B model and its SheetSage2 audio encoder, organized into ComfyUI checkpoints and audio encoder folders. The repository links to the original m-a-p/YuE2-3B and m-a-p/SheetSage2 model pages and is currently trending #30 on Hugging Face.

Hugging Face trending models · 6d agoModel release 2 sources

HarnessVLN: Unifying Training-Free Embodied Navigation through an Agent Harness

Researchers introduce HarnessVLN, a zero-shot training-free agent harness that sets new training-free SOTA on vision-language navigation benchmarks including R2R and HM3D.

HarnessVLN is a zero-shot, training-free framework for embodied vision-language navigation that coordinates perception, retrieval, grounding, navigation, recovery, and termination through a unified tool interface. It validates planner proposals against spatial evidence, geometric feasibility, and subgoal consistency, using hierarchical event memory and a persistent Spatiotemporal Graph that stores reusable spatial evidence and failure annotations. It reports success rates of 60.8% on R2R, 53.9% on RxR, 76.0% on HM3D-v2, and 59.3% on HM3D-OVON, surpassing prior training-free state of the art, with real-world humanoid deployment demonstrated.

Hugging Face daily papers · 3d agoAI research

Adobe fixes critical Magento zero-day exploited to backdoor servers

Adobe emergency-patches actively exploited max-severity Magento/Adobe Commerce zero-day CVE-2026-75650 (StyleSmuggler), used since Sept 4 to backdoor servers.

Adobe released an emergency hotfix (VULN-39341) for CVE-2026-75650, a max-severity zero-day dubbed StyleSmuggler affecting Adobe Commerce 2.4.4-2.4.9, Adobe Commerce B2B 1.3.3-1.5.3, and Magento Open Source 2.4.6-2.4.9, enabling arbitrary code execution. Sansec reports the flaw has been exploited since at least September 4 to plant a backdoor whose C2 host is disguised as an NTP server, leaving traces like 'Payment Transaction Failed Reminder' emails. A second attacker with unrelated tooling is exploiting the flaw to deploy a 485-byte PHP web shell that collects server details, checks pub/media writability, and exfiltrates data to an oast.site subdomain. Adobe recommends immediate hotpatching plus rotation of all secrets including admin passwords, API keys, database credentials, and SSH keys.

BleepingComputer · 8d agoExploit / PoC in the wildCVE-2026-756501

Show-Harness: Just a VLM Agent Can Play Robots

Show-Harness lets VLM agents control robots via discrete semantic action units, outperforming VLA baselines zero-shot and after light fine-tuning.

Show-Harness is an embodied agent harness that exposes discrete semantic action units a VLM reasons over, with embodiment-specific interpreters grounding them into local robot actions. It enables zero-shot robot control with closed-source frontier VLMs and low-cost adaptation of small open-source VLMs using only a few GPU-hours of fine-tuning. The companion GUMI (GUI Manipulation Interface) extends the same semantic action space to GUI-based demonstration collection without specialized teleoperation hardware. Experiments show robust generalization across tasks, embodiments, and environments, beating representative agentic and VLA paradigms.

arXiv cs.AI / cs.LG / cs.CL · 7d agoAI research

Can your coding style predict whether your code is vulnerable?

University of Massachusetts Dartmouth researchers present VulStyle, a stylometry-based vulnerability detector that also exposes benchmark reliability problems.

VulStyle combines stylometric features with syntax-tree structure and source tokens, pre-trained on about 4.9 million functions across seven programming languages and fine-tuned on five vulnerability detection datasets. It beat token-only detectors on some benchmarks but its F1 drops sharply on DiverseVul, which the authors link to noisy labels inflating reported performance across popular datasets. The authors argue style-aware detection should be harder to evade but did not test this empirically, and they note that uniform LLM-generated code may strip away the individual developer style the model depends on.

Help Net Security · 24d agoResearch1

Can LLMs Engineer Their Own Agent Harness? ByteDance Seed’s HarnessDev Says Only 34 of 64 Changes Generalize

ByteDance Seed's HarnessDev benchmark finds LLM-built agent harnesses trail human engineering on code and search, with only 34 of 64 revisions generalizing.

Researchers from ByteDance Seed, SUTD, Georgia Tech, M-A-P, and TokenWave.AI introduce HarnessDev, a benchmark that evaluates the runnable agent harness an LLM writes rather than its answers, using Creation and Evolution stages across SWE-bench Pro, Terminal-Bench 2.1, MLE-bench, EQ-Bench3, and BrowseComp (2,207 instances). Six creator models including Opus 4.8, GPT-5.5, Gemini 3.1 Pro, DeepSeek V4 Pro, Qwen 3.7 Max, and Seed 2.0 Pro were tested; Opus 4.8 posted the best average of 67.8 versus an 86.2 human-engineered reference. Self-built harnesses beat references on writing and ML experimentation but lag badly on code and search, and quality proved executor-specific: Opus 4.8 fell from 69.3 to 33.0 on SWE-bench Pro when the executor was switched to Gemini. Evolution gains were small and noisy: of 64 adjacent changes, feedback and held-out scores agreed only 34 times (53.1%), and much generated state and memory code never executed.

MarkTechPostupdated · 12h agofirst · 5d agoAI research 19 sources

microsoft/VibeVoice-ASR-Streaming-7B — new model trending #27 on Hugging Face

Microsoft released VibeVoice-ASR-Streaming-7B, an open streaming ASR model with speaker attribution, custom hotwords, and support for 10 languages under MIT license.

Microsoft Research released VibeVoice-ASR-Streaming-7B on Hugging Face, a unified streaming speech recognition model that continuously transcribes who said what as speech arrives. The 7B model supports customized hotwords for domain-specific terms and 10 languages including Chinese, English, French, German, Italian, Japanese, Korean, Portuguese, Russian, and Spanish. Code is available at github.com/microsoft/VibeVoice with a live demo, and a technical report is on arXiv (2609.02812). The model is licensed under MIT.

Hugging Face trending models · 14d agoModel release1

Alissonerdx/Minimax-H3-ComfyUI — new model trending #30 on Hugging Face

Alissonerdx released a rank-64 LoRA sharpening videos from MiniMax H3 in ComfyUI via guide-latent conditioning, trending #30 on Hugging Face.

Alissonerdx published Minimax-H3-ComfyUI, a repository containing a rank-64 LoRA (minimax_h3_lms_v1.0_r64) that sharpens source videos while keeping them photorealistic. It targets the MiniMax H3 ref2va base model, runs with Comfy-Org's ComfyUI-ready weights, and uses the native MiniMaxH3AddGuide node for aligned guide-latent conditioning. The release ships with a ComfyUI workflow, eight before/after example videos, and an Apache 2.0 license.

Hugging Face trending models · 10d agoModel release2

Rebuilding AUTOMATIC1111 with Gradio Workflow

Hugging Face demonstrates rebuilding the AUTOMATIC1111 Stable Diffusion web interface using its Gradio Workflow framework.

Hugging Face published a post showing how to rebuild the AUTOMATIC1111 Stable Diffusion WebUI experience with the Gradio Workflow framework. The article body was unavailable, so details beyond the title are limited, but the piece appears to be a tutorial on composing interactive AI interfaces with Gradio Workflow components.

Hugging Face Blog · 7d agoAI tools & infra

Multi-Vector (Late Interaction) Embedding Models with Sentence Transformers

Hugging Face details building and using multi-vector late-interaction embedding models with Sentence Transformers for retrieval workloads.

Hugging Face published a guide on multi-vector, late-interaction embedding models (ColBERT-style) supported through Sentence Transformers. The post covers how practitioners can build and use these models for retrieval and RAG pipelines. It is a developer tooling and technique write-up, not a security advisory.

Hugging Face Blog · Aug 18, 2026AI tools & infra1

TANGO: Humanoid Navigation in Cluttered Environments with a Whole-Body Vision-Language-Action Model

Researchers introduce TANGO, a whole-body vision-language-action model enabling zero-shot language-guided humanoid navigation on the Unitree G1 robot.

TANGO addresses humanoid navigation in cluttered indoor environments by predicting 29-DoF joint-space actions directly from natural-language instructions and egocentric RGB, rather than 2D path planning. It is trained entirely in simulation via a pipeline combining global path planning, kinematic whole-body motion generation, obstacle-aware motion editing, and RL-based tracking. In simulation it achieves state-of-the-art vision-language navigation performance and transfers zero-shot to a Unitree G1 humanoid without any real-world navigation data.

Hugging Face daily papers · 9d agoAI research1

Hackers Deploy New SloppyRAT via ClickFix to Enable Ransomware Lateral Movement

Zscaler ThreatLabz reports SloppyRAT, a new Windows RAT delivered via ClickFix lures that supports reconnaissance, reverse SOCKS proxying, and lateral movement for ransomware affiliates.

Zscaler ThreatLabz first observed SloppyRAT in June 2026, delivered through multi-stage ClickFix chains that abuse finger.exe, a renamed curl.exe, and IronPython to load a reflective Python payload. The C++ implant implements 47 PowerShell-like commands via native APIs and direct syscalls, modifies Microsoft Defender settings, and establishes a reverse SOCKS proxy that lets operators pivot into internal networks. It is linked to CastleLoader and CastleRAT components and uses certificate-pinned HTTPS C2 with an EtherHiding fallback via Polygon JSON-RPC. Zscaler assesses it as an intrusion-enablement tool for ransomware operations, though flawed persistence logic suggests the family is still under development.

GBHackersupdated · 5d agofirst · 6d agoMalware in the wild 3 sources

SynthGait-19K: A Physically Grounded Synthetic Video Dataset for Gait Parameter Estimation

Researchers release SynthGait-19K, a synthetic video dataset with 19,272 walking videos for training gait parameter estimation models.

SynthGait-19K is a physically grounded synthetic video dataset built from 6,427 MoCap sequences of 437 subjects, yielding 19,272 walking videos with SMPL motion and annotations for six gait parameters. The authors introduce Gait2Vid, a pipeline that unifies heterogeneous MoCap recordings and synthesizes RGB videos under controllable viewpoints, validating gait events against force-platform measurements. Using the dataset they benchmark direct RGB, pose-based, biomechanical, and human-mesh-recovery approaches, and introduce GaitXFormer as a direct RGB reference model. Findings show synthetic supervision transfers to real video, while spatial gait parameters are more sensitive to visual domain shift.

Hugging Face daily papers · 9d agoAI research

Novel macOS Infostealer AmnesiaStealer Spread via ClickFix

Novel macOS infostealer AmnesiaStealer spreads via ClickFix social engineering and lets attackers remotely control victims' browsers to steal cookie data.

Infosecurity Magazine reports on AmnesiaStealer, a novel macOS infostealer distributed through the ClickFix social engineering technique. The malware includes capabilities that let attackers take remote control of the victim's browser in order to steal cookie data. The novel remote-browser-control function distinguishes it from commodity macOS infostealers.

Infosecurity Magazine · Aug 14, 2026Malware in the wild

TANGO: Humanoid Navigation in Cluttered Environments with a Whole-Body Vision-Language-Action Model

Researchers present TANGO, a whole-body vision-language-action model enabling humanoid robots to traverse cluttered spaces from language instructions.

TANGO predicts 29-DoF joint-space actions from egocentric RGB observations and natural-language instructions for whole-body humanoid navigation, going beyond 2D path planning. It is trained entirely in simulation using global path planning, kinematic whole-body motion generation, obstacle-aware motion editing, and RL-based tracking. The model reports state-of-the-art simulation performance and was deployed zero-shot on a Unitree G1 humanoid without any real-world navigation training data.

arXiv cs.AI / cs.LG / cs.CL · 8d agoAI research1

HarnessTax: How Much Does the Harness Matter for Coding Agents?

HarnessTax is a research project measuring how much the harness, the scaffolding around LLMs, affects coding agent performance.

HarnessTax examines how much the harness — the scaffolding, prompts, and tooling wrapped around a large language model — contributes to coding agent results, as opposed to the underlying model itself. The project was posted on Hacker News on September 16, 2026, where it drew 42 points and 9 comments. Further details are available on the project's GitHub Pages site.

Hacker News · AI · 8h agoAI researchHN 42↑ · 9 comments

WarmBloodAban/Minimax-h3_Singularity — new model trending #22 on Hugging Face

Community fine-tune Minimax-h3_Singularity enhances MiniMax-H3 video generation with HDR quality, distant face restoration, and improved motion, trending #22 on Hugging Face.

Minimax-h3_Singularity is a community fusion fine-tune of the MiniMax-H3 multimodal video generation model, built from multiple checkpoints and refined with pruning and weight optimization. It supports Text-to-Video, Image-to-Video, Reference-to-Video, and Video-to-Video workflows in ComfyUI, and claims improvements in HDR clarity, distant face restoration, motion fluidity, and fantasy VFX. The authors recommend pairing it with the minimax_h3_ref2v_turbo_4step_v0.1 LoRA for four-step accelerated inference, and an online demo is available via RunningHub.

Hugging Face trending models · 11d agoModel release7· 1 read

TaichuAI/ZDTaichu5.0-9B — new model trending #30 on Hugging Face

TaichuAI released ZDTaichu5.0-9B, an open multimodal VLM built on Qwen3.5-9B targeting spatial reasoning, embodied AI, and agentic tool use.

TaichuAI released ZDTaichu5.0-9B, a multimodal vision-language model pairing a Qwen3.5-9B language decoder with a C-RADIOv4-H vision encoder, supporting text, single/multiple images, and video with any-resolution input and a 128K-token context. It introduces Entropy-Gated Adaptive Recurrent Reasoning, which allocates extra latent refinement steps to harder tokens. Reported benchmarks include 87.7 on TAU2-Bench, 71.4 on Claw-Eval, 93.7 on IFEval, 48 on ERQA, and 56 on RoboSpatial, leading compared 10B-scale open VLMs on agent and spatial tasks. The weights are available on Hugging Face, GitHub, and ModelScope, where it is trending at #30.

Hugging Face trending models · 13d agoModel release1

Fine-tuning a 350M Model for Better Structured Outputs in 100 GRPO Steps

Hugging Face guide fine-tunes a 350M-parameter model with 100 GRPO steps to improve structured output reliability.

A Hugging Face blog post demonstrates fine-tuning a 350M-parameter model using GRPO (Group Relative Policy Optimization) with TRL over 100 training steps. The stated goal is more reliable structured outputs from small language models. No article body was available, so details beyond the title are limited.

Hugging Face Blog · 14d agoAI tools & infra