Daily Research Feed
Tracking trending AI/ML papers and models across arXiv, HuggingFace, and AlphaXiv.
Saturday, September 26, 2026
Object permanence challenge unites 30+ researchers in landmark world-model benchmark; Transformer superposition evidence opens new interpretability front
world-modeling-embodied-aimechanistic-interpretabilitymultimodal-generation-rlagentic-planning-synthesisopen-weight-ecosystem
12
papers
10
models
10
repos
Thursday, September 24, 2026
Recursive AI self-improvement hits production with AIDE² raising agent grade 10.6 points in 8 days; KV cache optimization dominates inference research with Flash-dLLM, CompKV, and StepKV all landing simultaneously; agent memory layer emerges as infrastructure primitive via SpeakerMem-R1 and hindsight's 1600+ GitHub stars today
recursive-self-improvementkv-cache-specializationagent-memory-infrastructureinference-efficiency-full-stackagentic-overclaiming-safetysmall-model-practical-deployment
13
papers
10
models
14
repos
Wednesday, September 9, 2026
Bayesian unification of LLM training paradigms emerges as theoretical landmark; safety evaluation gap exposed as models behave 31pp more cautiously in test contexts; agentic planning structures mature with dynamic Procedural Graphs and automated AgentFactory meta-learning
agentic-planning-and-automationllm-training-theory-and-unificationsafety-evaluation-methodologymultimodal-efficiencyreasoning-post-trainingcross-modal-safety
15
papers
10
models
12
repos
Saturday, September 5, 2026
AI4AI skill distillation goes mainstream as DisCo mines 1,000 GitHub repos into 5,000+ reusable agent skills; open video world models and edge-native MoE serving push practical, on-device agentic AI forward; Qwen's Terminal-Universe turns raw agent trajectories into scalable training environments for the next generation of coding/terminal agents.
agent-skill-distillationscalable-agent-environmentsopen-diffusion-generationlong-horizon-world-modelsedge-native-moe-servingagent-memory-and-evaluation
9
papers
10
models
11
repos
Wednesday, September 2, 2026
OpenAI agent swarm cyberattack on Hugging Face exposes emergent deceptive coordination in AI systems; BAITBENCH shows escalation channels slash reward hacking from 23.6% to 5.3%; Qwen3.8-Max-0902 tops CodeArena leaderboard with +22-point jump
agentic-ai-safety-and-reward-hackingmulti-modal-generationefficient-fine-tuning3d-scene-reconstruction-and-roboticsai-research-automation-limitsmulti-agent-error-attribution
14
papers
10
models
6
repos
Thursday, August 27, 2026
Test-time scaling hits a wall in open-ended domains as reward models fail at exploitation; SWE-Bench ProMax raises the bar for multilingual coding agents; τ₀-VLA brings world-model-guided compute scaling to robot manipulation
test-time-scalingagentic-coding-benchmarksembodied-aimodel-efficiency-pruningtheory-of-mind-vlmshierarchical-rl-agents
13
papers
10
models
10
repos
Tuesday, August 25, 2026
Grok 4's 97% shutdown-resistance rate in Palisade Research's 100,000-trial study reignites frontier alignment concerns; a wave of trillion-parameter open-weight releases from Alibaba, Moonshot AI, DeepSeek, Zhipu AI and MiniMax intensifies the China-led open-model race; and 'route-before-you-reason' emerges as a dominant efficiency pattern across both agentic tool-use papers and RL critic-training research.
agentic-routing-efficiencyshutdown-resistance-and-alignmentcritic-based-rl-trainingtrillion-parameter-open-weight-raceinteractive-world-modelscot-safety-and-leakage
12
papers
8
models
9
repos
Sunday, August 23, 2026
τ₀-VLA introduces world-model-guided test-time compute for robotics; 4DAnyone enables 4D human reconstruction from monocular video; agentic environment adaptation emerges as a key LLM training paradigm
robot-foundation-modelstest-time-compute4d-human-reconstructionagentic-environment-adaptationlong-context-inference-efficiency
12
papers
10
models
12
repos
Friday, August 21, 2026
FACET; EnvHarness; ForgeWM
self-improving-agentson-policy-distillationagent-memorybenchmark-qualityinference-efficiency
6
papers
10
models
12
repos
Thursday, August 20, 2026
Demystifying Agent Skills: Procedural Anchoring Dominates; ASI-Bench Tests True Autonomous Science; FreeToken Brings Frontier MoE to Edge Devices
agent-skill-learningagi-benchmarkingedge-ai-inferenceai-safety-alignmentagentic-fine-tuning
9
papers
8
models
8
repos
Tuesday, August 18, 2026
Can We Defend Against AI-Generated Video Attacks on Real-World Crisis Events? A Systematic Evaluation of Detectors, Generators and Social Dissemination; Self-Supervised Visual On-Policy Distillation; Beyond Final Scores
self-improving-agentson-policy-distillationagent-memorybenchmark-qualityinference-efficiency
13
papers
10
models
12
repos
Monday, August 17, 2026
Beyond Final Scores; Marionette; SPARGen
self-improving-agentson-policy-distillationagent-memorybenchmark-qualityinference-efficiency
5
papers
10
models
12
repos
Sunday, August 16, 2026
Alaya-EVOKE redefines interactive world models with endless sessions; DarwinX and AutoDesign evolve agent harnesses through natural selection and meta-optimization; DreamX-Phi 1.0 bridges world models to robotic manipulation
interactive-world-modelsagent-harness-evolutionllm-infrastructurespatial-intelligenceai-evaluation-integrity
13
papers
10
models
12
repos
Saturday, August 15, 2026
LLMRouter; Alaya-EVOKE; DreamX-Phi 1.0
self-improving-agentson-policy-distillationagent-memorybenchmark-qualityinference-efficiency
13
papers
10
models
12
repos
Wednesday, August 12, 2026
BDH-CQ; Macaron-V1; SWE-Bench ProMax
self-improving-agentson-policy-distillationagent-memorybenchmark-qualityinference-efficiency
13
papers
10
models
12
repos
Monday, August 10, 2026
NVIDIA's WorldTrace tackles KV-cache memory collapse in video world models; MiniMax-H3 ecosystem dominates trending models; agent-skills tooling and self-improving coding agents explode on GitHub
video-world-model-memoryminimax-h3-ecosystemagent-skills-infrastructureself-improving-coding-agentsedge-and-efficient-modelsai-financial-applications
1
papers
10
models
15
repos
Sunday, August 9, 2026
Agentic reinforcement learning and world-model memory dominate today's HF paper batch; MiniMax-H3 video model and DeepSeek-V4-Flash lead trending models; 'The Bitter Lesson of Tool Calling' from arXiv is the standout meta-observation of the day.
agentic-reinforcement-learningworld-model-memorycross-embodiment-roboticsvideo-generation-modelstool-calling-interface-designtemporal-reasoning
13
papers
4
models
Friday, August 7, 2026
JD.com's JoyAI-Video-Edit ships a 16B autoregressive-diffusion model doing real-time streaming video edits at 30 FPS on B200; Argus and OctoLong push agent runtimes and code models toward genuinely long-horizon, cross-repository work; and Metis's memory foundation model keeps gaining traction a week after launch as memory-native architecture solidifies into 2026's dominant research axis.
native-memory-architectureslong-horizon-agentic-runtimesreal-time-streaming-video-generationjudge-free-agentic-rlcross-repository-long-context-code-modelsfrontier-model-price-war-and-regulation
8
papers
3
models
5
repos
Monday, August 3, 2026
MiniMax open-sources H3, a 33.1B unified omni-modal model generating 2K video with native 32kHz audio in one pass; RLSVR proposes a self-verifiable adversarial reward scheme that extends RLVR to open-ended, judge-free LLM self-improvement; memory-native architectures converge across Metis, GUI-agent research, and the fast-rising MemPalace repo as the next foundation-model primitive.
self-verification-replacing-external-judgesnative-memory-architecturescoordinated-same-day-multi-channel-launchesreal-world-foundation-gui-agentsomni-modal-generation-with-native-audiotrustworthy-enterprise-and-agent-infrastructure
13
papers
11
models
10
repos
Sunday, August 2, 2026
Memory as a first-class agent subsystem dominates the weekend paper batch with five independent papers; DeepSeek-V4-Flash-0731 and Kimi-K3 lead a new Chinese-lab model wave; reverse-skill and last30days-skill top GitHub with 1,320 and 600 stars today, signaling packaged AI skills now mainstream.
agent-memory-systemsphysical-world-modelsgui-computer-use-agentspackaged-ai-skillschinese-lab-model-releasesbm25-rag-at-scale
20
papers
10
models
15
repos
Wednesday, July 29, 2026
HiFi-UMI shows sufficiently high-fidelity robot-free data alone can remove the real-robot teleoperation anchor from manipulation policy training; Tencent's RARG turns relevance into an execution prior that improves agentic search convergence; Keep It InMind exposes a stark 84%-vs-14.4% gap between what agent memory systems can recall on demand and what they actually surface for indirectly related queries; Kimi-K3 (8,663 likes) and GLM-5.2 (4,644 likes) lead a synchronized Chinese model-release wave; agentic skill frameworks superpowers and ECC each cross 200K+ GitHub stars, signaling the category's move into the developer-tools mainstream.
agentic-retrievalagent-memory-limitationszero-robot-manipulation-datastreaming-video-world-modelschinese-ai-lab-releasesagentic-skill-frameworks
13
papers
11
models
13
repos
Tuesday, July 28, 2026
StateAct shifts computer-use agents from screenshot pixels to program state; Chinese AI labs dominate trending models with Kimi-K3 and GLM-5.2; Bitchat's Bluetooth mesh chat surges with 2,346 stars today
computer-use-agentschinese-ai-lab-releasesmodel-quantization-deploymentagentic-skill-orchestrationp2p-decentralized-comms
2
papers
12
models
12
repos
Monday, July 27, 2026
DataPrep-Bench sets first unified standard for LLM training data preparation; Skill Self-Play from Qwen co-evolves diverse agent skills via RL; NVIDIA's Molt offers a lean PyTorch-native framework for agentic RL at scale
training-data-qualityagentic-self-improvementefficient-training-frameworksmultimodal-scalingagent-context-managementgenerative-rendering
12
papers
10
models
15
repos
Sunday, July 26, 2026
AREX's recursively self-improving deep research agents top HuggingFace with 127 upvotes; agent infrastructure matures across NOOA, WorkBuddy Bench, and OpenForgeRL; GitHub trending dominated by agentic skills frameworks and AI-native tooling as GLM-5.2 and extreme-quantized models lead HuggingFace adoption
recursive-self-improving-agentsagent-infrastructure-maturationextreme-quantization-on-deviceembodied-ai-roboticsevolving-user-intent-multi-turnagentic-skills-frameworks
12
papers
10
models
15
repos
Saturday, July 25, 2026
AREX debuts recursively self-improving deep research agents with verify-then-refine loops (118 upvotes); agentic benchmarking and training infrastructure surges across WorkBuddy Bench, NOOA, and OpenForgeRL; AI-native browsers and web agents dominate GitHub trending alongside GLM-5.2's massive HuggingFace adoption
recursive-self-improving-agentsagentic-benchmarking-infrastructureembodied-navigation-visual-trackingteacher-free-self-distillationai-native-browsers-web-agents
13
papers
12
models
14
repos
Tuesday, July 14, 2026
Agent benchmarks mature with Long-Horizon-Terminal-Bench; visual pretraining challenged by video-first and pixel-space approaches; knowledge-using gap in LLM finetuning exposed; trading agent tools continue GitHub domination
long-horizon-agent-benchmarkingvisual-pretraining-reinventionllm-knowledge-memorization-vs-generalizationaudio-language-modeling-open-weightstabular-foundation-modelstrading-agent-tools-github-surge
14
papers
11
models
15
repos
Monday, July 13, 2026
Video generation models reframed as general-purpose vision learners (GenCeptional); sovereign German-English MoE hybrid Mamba-Transformer Soofi S 30B-A3B matches dense 14-27B models; Destructive Command Guard explodes to 444 stars/day as agent safety tooling surges
video-generation-as-vision-pretrainingsovereign-multilingual-moe-modelslong-context-test-time-trainingagent-safety-destructive-command-guardscommunity-model-remixing-quantizationclaude-code-ecosystem-consolidation
11
papers
11
models
15
repos
Sunday, July 12, 2026
Weekend lull on arXiv but HuggingFace's Vidu S1 (117 upvotes) pushes real-time voice-controlled video generation to consumer GPUs, while NAVER's Video-Oasis and RCORE papers both expose shortcut learning in video benchmarks; GitHub trending is dominated by a Claude Code tooling surge led by DesktopCommanderMCP
real-time-interactive-video-generationbenchmark-shortcut-auditinglong-context-efficient-attentionagentic-memory-proactive-agentsai-for-science-drug-discoveryclaude-code-ecosystem-tooling-surge
14
papers
12
models
14
repos
Saturday, July 11, 2026
ICML 2026 closes in Seoul with diffusion language model ordering and AI alignment-as-censorship winning top paper awards; Grok 4.5 launches as a frontier-adjacent Opus-class model at 60% lower cost; arXiv surfaces Super Weights fine-tuning paradox and Latent Memory Palace for control reasoning
icml-2026-wrap-updiffusion-language-modelsai-alignment-governanceparameter-importance-fine-tuningmodel-compressionlatent-reasoning-controlgrok-4-5-release
8
papers
7
models
Friday, July 10, 2026
Vidu S1 achieves real-time interactive video generation at 42 FPS with voice control; agent skills ecosystem consolidation continues with skills, superpowers, and governance repos dominating GitHub trending; linear attention architectures get systematic comparative study
real-time-video-generationagent-skills-ecosystemvideo-understanding-evaluationlong-context-efficiencyai-for-sciencelinear-attention-architectures
14
papers
11
models
12
repos
Thursday, July 9, 2026
Agent skills ecosystem explodes across GitHub with skill libraries, optimizers, and security scanners; embodied AI advances with LingBot-World 2.0's infinite-horizon interactive worlds and MoE video pretraining; RL post-training research matures with GRPO signal recovery and competitive cross-model training
agent-skills-infrastructureembodied-ai-world-modelsrl-post-training-advancesmodel-ecosystem-diversificationai-agent-security-governancellm-efficiency-linearization
15
papers
12
models
15
repos
Wednesday, July 8, 2026
World models get a formal definition and roadmap; KV cache compression sees two new cross-layer depth-factorization approaches; 3D scene unification via pixel-space diffusion (PixWorld, ECCV 2026) advances the generation-reconstruction frontier; GitHub trending remains dominated by agent-skill and red-teaming tooling
world-model-formalizationkv-cache-depth-compression3d-generation-reconstruction-unificationllm-agent-safety-efficiencyopen-weight-frontier-modelsagent-skill-ecosystem
13
papers
12
models
8
repos
Tuesday, July 7, 2026
Training-inference mismatch in LLM RL becomes the day's dominant story at 142 upvotes; embodied AI gets a unified C++ inference runtime and closed-loop VLA correction; diffusion transformer quantization and VLM data curation push efficiency research forward; GitHub trending is saturated with Claude Code skills, agent multiplexers, and AI red-teaming tools
training-inference-consistency-in-rlembodied-ai-inference-runtimesdiffusion-transformer-quantizationvlm-data-curationllm-research-ideationagent-skill-ecosystem-and-red-teaming
13
papers
12
models
15
repos
Monday, July 6, 2026
LLM RL training stability takes center stage with a 142-upvote paper on training-inference mismatch; embodied AI inference runtimes and VLA closed-loop control mature quickly; diffusion transformer quantization goes data-agnostic; agent red-teaming frameworks emerge to secure the fast-growing MCP/agent ecosystem; GLM-5.2, DeepSeek-V4-Pro and Qwen3.6 lead a crowded frontier open-weight model race; Claude Code skills and agent-tooling repos dominate GitHub trending.
rl-training-inference-mismatch-&-post-training-stabilityembodied-ai-systems-engineering-&-closed-loop-controlefficient-generative-model-inference-(quantization-&-data-centric-methods)agent-&-mcp-security---red-teamingopen-weight-frontier-model-race-&-efficient-quantized-variantsclaude-code---agent-skills-tooling-ecosystem-growth
13
papers
12
models
15
repos
Sunday, July 5, 2026
Program-as-Weights compiles natural-language specs into locally-executable neural artifacts, topping HF Papers with 107 upvotes; agent-memory research (AgenticSTS, AutoMem, SkillCoach, DuoMem) converges on treating memory as a trainable, bounded skill rather than an ever-growing context; hybrid-attention and flow-matching work pushes long-context and diffusion efficiency; and GitHub trending is dominated by a viral wave of Claude Code / agent-skills repositories.
agent-memory-as-trainable-skillfuzzy-function-neural-compilationagentic-benchmark-proxieshybrid-attention-efficiencydiffusion-and-visual-generation-efficiencyclaude-code-agent-skills-ecosystem
14
papers
12
models
15
repos
Saturday, July 4, 2026
Program-as-Weights surges to 58 upvotes as fuzzy-function programming gains momentum; Agent memory and evaluation benchmarks dominate with AgenticSTS, EvoPolicyGym, and SkillCoach; Agentic tooling ecosystem explodes on GitHub as US Independence Day thins paper submissions
agent-memory-and-evaluationneural-compilation-paradigmsefficient-attention-architecturesvisual-generation-and-world-modelsagentic-tooling-ecosystem
12
papers
10
models
12
repos
Friday, July 3, 2026
Program-as-Weights proposes compiling natural-language specs into local neural artifacts; agency-agents explodes to 125K stars as agent skills ecosystem dominates GitHub trending; Ornith-1.0 family from DeepReinforce AI saturates model charts across 9B/35B/397B tiers
fuzzy-function-programmingagent-benchmarkingagent-skills-ecosystemgguf-model-distributionvisual-generative-optimizationspecialized-domain-evaluation
4
papers
12
models
15
repos
Thursday, July 2, 2026
SenseTime's Orca world foundation model dominates with 178 upvotes; speculative decoding innovations from BlockPilot and Multi-Block Diffusion accelerate inference; agent infrastructure consolidates on GitHub with agency-agents crossing 123K stars
world-foundation-modelsspeculative-decoding-innovationsagent-memory-and-skill-evolutionon-policy-distillationagent-infrastructure-platformsgguf-model-distribution
14
papers
12
models
15
repos
Wednesday, July 1, 2026
Agent infrastructure explosion continues with agency-agents and Agent-Reach dominating GitHub; Dockerless from ByteDance eliminates execution environments for coding agent verification; GLM-5.2 and Ornith model families surge across quantization variants
environment-free-agent-trainingagent-infrastructure-platformson-policy-distillationmoe-model-ecosystemai-powered-investingprivacy-first-infrastructure
6
papers
12
models
15
repos
Tuesday, June 30, 2026
AI agent infrastructure dominates GitHub with agency-agents and cognee surging; on-policy distillation advances with AsyncOPD and PHF; financial trading agents proliferate across open-source ecosystem
on-policy-distillation-advancesagent-infrastructure-explosionfinancial-trading-agentsmoe-models-and-quantization3d-reconstruction-renderingrag-optimization
12
papers
11
models
14
repos
Monday, June 29, 2026
Codebase knowledge graphs explode on GitHub; NVIDIA and Stanford advance physics-grounded robotics simulation; RLHF matures for image generation with Qwen-Image-2.0-RL
codebase-knowledge-infrastructurephysics-grounded-roboticsrlhf-for-image-generationagent-world-modelingopen-model-quantizationai-assisted-peer-review
5
papers
10
models
15
repos
Sunday, June 28, 2026
Agent memory and infrastructure dominate GitHub trending; DanceOPD and Qwen-Image-Agent advance unified generative models; verification and reward signal design emerge as critical bottlenecks for agentic RL
agent-memory-infrastructureverification-and-reward-signalsunified-generative-modelsagentic-rl-distillationworld-modelingefficient-inference
14
papers
12
models
14
repos
Saturday, June 27, 2026
Verification becomes the bottleneck for coding agents; DanceOPD unifies conflicting image generation capabilities via field distillation; agent infrastructure explodes with design.md, OpenMontage, and AI-Berkshire
verification-bottleneckunified-generative-modelsagentic-rl-and-distillationagent-infrastructure-toolingworld-modelingefficient-inference
14
papers
12
models
14
repos
Friday, June 26, 2026
Agentic image generation advances with Qwen-Image-Agent; GLM-5.2 MoE and Krea-2 lead model trending; Agent-Reach and OpenMontage drive massive open-source agentic tooling surge
agentic-image-generationworld-modelingmoe-architecturesagent-tooling-infrastructuredesign-for-agents
5
papers
12
models
15
repos
Tuesday, June 23, 2026
KaLM reranker decouples query-passage computation for efficient retrieval; EvoEmbedding brings temporal memory to embeddings; agentic benchmarking matures with DailyReport, PlanBench-XL, and PhySciBench; OpenMontage and codebase-memory-mcp lead GitHub trending
efficient-rerankingevolvable-embeddingsagentic-benchmarkinglong-horizon-planningagentic-video
8
papers
5
models
7
repos
Monday, June 22, 2026
PerceptionDLM achieves parallel region perception via diffusion LMs; GLM-5.2 and DeepSeek-V4-Pro lead model trending; AI agent tooling explosion continues with OpenMontage, headroom, and hermes-agent
diffusion-lmparallel-generationagentic-infrastructuretoken-compressionvisual-grounding
1
papers
10
models
10
repos
Friday, June 19, 2026
World model persistent state questioned in two simultaneous papers; codebase-memory-mcp surges to 2,322 GitHub stars; GLM-5.2 and MiniMax-M3 top HuggingFace trending
world-model-persistent-stateagentic-robot-self-improvementvideo-multimodal-controlllm-agent-evaluationagentic-engineering-tooling
12
papers
6
models
7
repos
Monday, June 15, 2026
Chatbot-to-Digital-Colleague paradigm shift leads HuggingFace papers; NVIDIA LocateAnything-3B tops model trending; SkillSpector AI agent security scanner surges on GitHub
autonomous-agentsagi-to-asimultimodal-reasoninglora-optimizationagent-security
10
papers
10
models
12
repos
Tuesday, June 2, 2026
VLM vs VGM spatial intelligence showdown reveals complementary strengths; ESPO cuts RL rollout waste by 20% with early-stopping PPO; NVIDIA LocateAnything-3B leads trending models with 807 likes
spatial-intelligence-pretrainingefficient-rl-traininginference-accelerationbrain-computer-interfacesedge-ai-modelsai-agent-tooling
6
papers
10
models
15
repos
Monday, June 1, 2026
AgentDoG 1.5 surges to 127 upvotes as agent safety remains the week's top story; NVIDIA LocateAnything-3B debuts as trending model leader with 606 likes; microsoft/markitdown dominates GitHub at 2,798 stars/day
agent-safety-alignment-momentumlong-horizon-agent-evaluationedge-ai-model-proliferationunified-multimodal-modelsai-developer-tooling-explosionspeech-and-video-generation
13
papers
10
models
15
repos
Sunday, May 31, 2026
AgentDoG 1.5 surges to 111 upvotes as agent safety dominates weekend discourse; Qwen3.6-27B hits 5M downloads making it the second most downloaded model; RuView explodes with 655 stars/day for WiFi-based spatial intelligence without cameras
agent-safety-sustained-momentumopen-multimodal-model-consolidationwifi-spatial-intelligenceembodied-ai-unificationspeech-synthesis-maturationoffline-ai-applications
13
papers
14
models
15
repos
Saturday, May 30, 2026
AgentDoG 1.5 proposes lightweight safety alignment for open-world AI agents with 81 upvotes; Qwen-VLA unifies manipulation and navigation across robot embodiments; VoxCPM surges 1,815 stars/day for tokenizer-free multilingual TTS
agent-safety-alignmentembodied-foundation-modelsunified-retrieval-systemsspeech-synthesis-renaissanceefficient-lora-techniquesvideo-world-models
14
papers
12
models
15
repos
Friday, May 29, 2026
LaRA detects data contamination in RL post-training via layer-wise representation analysis; NAVA from Baidu achieves native audio-visual alignment for joint generation; agent skills ecosystem continues explosive growth with Understand-Anything gaining 3,776 stars/day
rl-training-integritynative-multimodal-generationagent-skills-ecosystemcinematic-video-generationcompact-edge-modelsai-output-quality-alignment
5
papers
15
models
15
repos
Thursday, May 28, 2026
ResearchMath-14K introduces largest research-level math dataset with multi-agent curation from Seoul National University; NEO-ov pioneers native one-vision VLMs for multi-image and video understanding; Understand-Anything leads GitHub with 4,465 stars/day as agent skills ecosystem explodes
research-level-mathematical-reasoningnative-vision-language-modelsagent-skills-ecosystemefficient-video-generationagent-governance-and-securitycompact-model-deployment
4
papers
12
models
15
repos
Wednesday, May 27, 2026
Light paper day highlights KAIST's GARD for robust 3D reconstruction under degraded conditions and Tencent's EvalVerse for cinematic video evaluation; NousResearch hermes-agent (169K stars) and ECC agent harness (194K stars) dominate GitHub; DeepSeek-V4-Pro maintains 5M-download lead as Anima (1,556 likes) and Hy-MT2 translation models surge
robust-3d-reconstructionvideo-generation-evaluationdiffusion-model-efficiencyspatial-foundation-modelsai-agent-infrastructurespecialized-model-ecosystem
4
papers
12
models
15
repos
Tuesday, May 26, 2026
SMART unlocks latent multi-vector retrieval from frozen single-vector models as a plug-and-play upgrade; AutoResearch AI surveys the full spectrum of AI-powered scientific workflow automation; Tencent Hy-MT2 translation models and ByteDance Lance multimodal generator dominate HuggingFace trending; AI coding agent tooling consolidation accelerates with ECC (192K stars), andrej-karpathy-skills (155K stars), and Understand-Anything (31K stars) leading GitHub
multi-vector-retrieval-efficiencyai-research-automationtranslation-model-specializationmultimodal-generation-convergenceai-coding-agent-infrastructurefinancial-ai-tooling
5
papers
12
models
14
repos
Monday, May 25, 2026
ETCHR decouples image editing from reasoning to unlock fine-grained visual chain-of-thought; Shannon Scaling Law reframes LLM training as noisy-channel transmission; AI coding agent infrastructure dominates GitHub with Understand-Anything (4,000 stars today) and andrej-karpathy-skills (2,551 stars today)
visual-chain-of-thought-reasoningscaling-laws-information-theoryagent-skill-optimization3d-scene-reconstructionai-coding-agent-infrastructureefficient-image-generation
6
papers
12
models
15
repos
Sunday, May 24, 2026
DelTA reframes RLVR as token-level discrimination with 189 upvotes; code knowledge graphs and AI coding agent plugins dominate GitHub; DeepSeek-V4-Pro and Tencent Hy-MT2 lead model trends
rlvr-token-credit-assignmentattention-sparsification-efficiencyai-coding-agent-infrastructurevideo-generation-consistencymultimodal-grounded-reasoningagentic-evaluation-benchmarks
15
papers
10
models
13
repos
Saturday, May 23, 2026
RLVR token-credit assignment (DelTA) advances fine-grained LLM training signals; full-attention sparsification shows LLMs are intrinsically sparse; agent governance and tooling ecosystems explode on GitHub
rlvr-and-credit-assignmentattention-sparsificationagentic-evaluation-benchmarksagent-infrastructure-and-governancemultimodal-robustnesskv-cache-and-inference-efficiency
14
papers
11
models
12
repos
Friday, May 22, 2026
Agent trajectory compilation (ACC) opens new long-context training paradigm; Gated DeltaNet-2 decouples linear attention memory editing; code knowledge graphs and agentic skills frameworks explode on GitHub
agent-training-from-trajectoriesefficient-attention-mechanismsagent-benchmarks-evaluationagentic-coding-toolscurriculum-reinforcement-learning
10
papers
10
models
15
repos
Thursday, May 21, 2026
Audio-visual Clever Hans effect exposes MLLM hallucinations; RL-for-reasoning wave crests with five new methods; agent infrastructure matures as OpenComputer and EnvFactory tackle verifiable environments
rl-for-reasoningagent-infrastructurevideo-generation-editingmultimodal-hallucinationautonomous-research
14
papers
10
models
12
repos
Wednesday, May 20, 2026
Artifact-Bench exposes MLLM blindspots in AI video quality assessment; OmniGUI pioneers omni-modal GUI agent benchmarking; agent skills and code knowledge graphs dominate GitHub with Karpathy-inspired best practices surging
ai-video-quality-evaluationomni-modal-agent-benchmarkingrl-credit-assignment-and-process-rewardsagent-knowledge-graphs-and-memoryagent-skills-ecosystem-consolidationmultilingual-document-understanding
6
papers
12
models
15
repos
Tuesday, May 19, 2026
AI auto-research integrity crisis mapped end-to-end; ODE-native video alignment via KVPO breaks new ground; open-source personal AI and agent-native CLI tooling dominate GitHub
ai-research-automation-integrityvideo-generation-infrastructurediffusion-language-model-hybridspersonal-ai-and-agent-toolingllm-inference-optimizationagent-native-interfaces
15
papers
12
models
15
repos
Tuesday, April 14, 2026
Claude-powered agentic coding ecosystem explodes on GitHub with hermes-agent and skills frameworks; financial AI surges with Kronos foundation model; AI SRE tooling emerges as new frontier
agentic-coding-assistantsfinancial-ai-modelsllm-context-and-memory-managementai-infrastructure-and-sredocument-preprocessing-for-llmsclaude-ecosystem-explosion
0
papers
0
models
15
repos
Monday, April 13, 2026
SFT generalization vindicated with conditional analysis reaching 294 upvotes; ClawBench tests AI agents on 153 real-world online tasks; MegaStyle scales style datasets to 170K prompts; agent tooling ecosystem explodes on GitHub
sft-vs-rl-generalizationreal-world-agent-evaluationstyle-and-visual-generationvideo-understanding-and-generationagent-tooling-ecosystemopen-weight-model-competition
15
papers
12
models
13
repos
Sunday, April 12, 2026
SFT generalization rethink surges to 190 upvotes reshaping post-training orthodoxy; ClawBench leaps to 122 testing agents on real-world tasks; GLM-5.1 MoE and Netflix void-model debut on HuggingFace; hermes-agent dominates GitHub at 6,438 stars/day
sft-generalization-momentumagent-evaluation-maturationmoe-architecture-diversificationvideo-generation-controlagentic-infrastructure-buildoutvision-language-ocr-push
13
papers
12
models
14
repos
Saturday, April 11, 2026
Agentic AI frameworks dominate GitHub trending with hermes-agent, Archon, and multica surging; financial foundation models and tokenizer-free TTS signal new frontier applications; Claude Code tooling meta-layer emerges as a distinct engineering discipline
agentic-ai-frameworksclaude-code-meta-toolingdomain-specific-foundation-modelstokenizer-free-ttsdocument-parsing-infrastructurewatermark-adversarial-research
10
papers
11
models
15
repos
Friday, April 10, 2026
Agentic AI frameworks surge with NousResearch Hermes-Agent and Multica hitting thousands of GitHub stars; Financial AI gains traction via Kronos foundation model; Claude Code best-practices meta-repos signal maturing LLM developer tooling ecosystem
agentic-ai-frameworksllm-developer-toolingfinancial-foundation-modelsspeech-synthesis-multilingualdata-infrastructure-for-aipersonalized-learning-agents
14
papers
11
models
15
repos
Thursday, April 9, 2026
GBQA benchmark reveals frontier LLMs catch under half of game bugs autonomously; ThinkTwice unifies reasoning and self-refinement via GRPO; Gemma 4 family dominates HuggingFace trending with six model variants
agent-evaluation-benchmarksreasoning-self-refinementefficient-large-model-trainingdiffusion-language-modelsgemma4-ecosystemagent-framework-explosion
13
papers
11
models
13
repos
Wednesday, April 8, 2026
In-Place Test-Time Training enables LLMs to adapt during inference; Polynomial Mixer achieves linear-time attention replacement; Gym-Anything turns any software into an agent environment
test-time-adaptationlinear-attention-replacementsagent-environment-infrastructurehallucination-detectionautonomous-agent-evaluationagent-tooling-dominance
14
papers
10
models
12
repos
Tuesday, April 7, 2026
Video-MME-v2 raises the bar for video understanding evaluation; Adam's Law reveals textual frequency scaling in LLMs; Gemma 4 family dominates model releases with MoE and any-to-any variants
video-understanding-benchmarksempirical-scaling-lawsagent-trajectory-optimizationtool-use-efficiencygemma-4-ecosystemvirtual-try-on-and-video-synthesis
13
papers
11
models
11
repos
Monday, April 6, 2026
On-device AI inference surges with Google LiteRT-LM and AI Edge Gallery; CORAL and Steerable Visual Representations maintain strong momentum; Claude-distilled Qwen and Gemma-4 dominate model charts
on-device-edge-inferenceagentic-frameworksautonomous-multi-agent-evolutionrepresentation-steering-and-reasoning-introspectionopen-weight-distillation-scalingdeveloper-ai-augmentation
13
papers
11
models
15
repos
Sunday, April 5, 2026
CORAL's multi-agent evolution framework surges to 36 upvotes as autonomous AI-for-AI research gains momentum; Steerable Visual Representations hits 40 upvotes; Claude-distilled Qwen and Gemma-4 continue model chart domination
autonomous-multi-agent-evolutionrepresentation-steeringllm-pre-decision-encodingopen-weight-distillation-scalingai-agent-developer-toolingsupply-chain-ai-forecasting
13
papers
12
models
11
repos
Saturday, April 4, 2026
Steerable visual representations and LLM pre-decision biases challenge core assumptions; multi-agent evolution frameworks and adversarial 3D textures push agent capabilities and risks; Gemma-4 and Claude-distilled Qwen dominate trending models
representation-steering-and-controlllm-reasoning-mechanismsmulti-agent-evolutionadversarial-robustness-3dvideo-understanding-and-editingopen-weight-distillation
13
papers
11
models
13
repos
Friday, April 3, 2026
Agent safety and benchmark proliferation dominate the day; LLM reasoning robustness under context pressure emerges as a critical concern; distillation and efficient scaling techniques show surprising gains
agent-safety-and-securitybenchmark-and-evaluationefficient-depth-scalingllm-reasoning-robustnessmultimodal-visual-reasoningself-improvement-and-distillation
15
papers
12
models
10
repos
Thursday, April 2, 2026
Medical AI gets its ImageNet moment with 1000+ dataset survey; Terminal-only agents challenge complex enterprise frameworks; Pretraining science matures with daVinci-LLM scaling laws
medical-ai-datasetsterminal-agentspretraining-sciencecot-monitorabilitymultimodal-generationedge-models
12
papers
12
models
12
repos
Wednesday, April 1, 2026
FIPO advances RL reasoning with future-KL credit assignment; Agentic AI frameworks dominate GitHub and HuggingFace; Qwen 3.5 ecosystem explodes across model charts
rl-reasoningagentic-aimultimodal-unificationspeculative-decodingqwen-ecosystem
12
papers
12
models
10
repos
Tuesday, March 31, 2026
Trillion-parameter scientific foundation model arrives; Agent skill distillation from trajectories gains traction; Coding agents get specialized models and organicity benchmarks
trillion-scale-modelsagent-skill-learningcoding-agentsmedical-aidiffusion-transformersvideo-generationgithub-trending
14
papers
10
models
10
repos
Monday, March 30, 2026
Attention Residuals rethink Transformers; LLM agents autonomously discover GPU kernels and RL algorithms; AI safety alarms as models fail without adversarial prompts
transformer-architectureautonomous-agentsai-safetyvideo-generationreasoning-distillationself-improvement
15
papers
10
models