VibeServe demonstrates that AI agents can synthesize bespoke LLM serving systems end-to-end, remaining competitive with vLLM in standard settings while outperforming it in six non-standard scenarios involving unusual models, workloads, or hardware.
hub
SGLang: Efficient Execution of Structured Language Model Programs
31 Pith papers cite this work. Polarity classification is still indexing.
31
Pith papers citing it
abstract
Large language models (LLMs) are increasingly used for complex tasks that require multiple generation calls, advanced prompting techniques, control flow, and structured inputs/outputs. However, efficient systems are lacking for programming and executing these applications. We introduce SGLang, a system for efficient execution of complex language model programs. SGLang consists of a frontend language and a runtime. The frontend simplifies programming with primitives for generation and parallelism control. The runtime accelerates execution with novel optimizations like RadixAttention for KV cache reuse and compressed finite state machines for faster structured output decoding. Experiments show that SGLang achieves up to 6.4x higher throughput compared to state-of-the-art inference systems on various large language and multi-modal models on tasks including agent control, logical reasoning, few-shot learning benchmarks, JSON decoding, retrieval-augmented generation pipelines, and multi-turn chat. The code is publicly available at https://github.com/sgl-project/sglang
hub tools
representative citing papers
NCCLZ decouples quantization and entropy coding across NCCL stack layers to enable overlapped compression, delivering up to 9.65x speedup over plain NCCL on scientific and training workloads.
A new image-bank harness and closed-loop on-policy data evolution method raises multimodal agent performance on visual search benchmarks from 24.9% to 39.0% for an 8B model and from 30.6% to 41.5% for a 30B model.
EEP makes wide expert-parallel MoE serving survive single-rank failures with an 11s recovery pause, 8s reintegration pause, and throughput restored to 95% of pre-fault level within 52s while staying within 4.4% of a fixed-membership baseline in steady state.
Dooly reduces LLM inference profiling costs by 56.4% via configuration-agnostic taint-based labeling and selective database reuse, delivering simulation accuracy within 5% MAPE for TTFT and 8% for TPOT across 12 models.
Hosted open-weight LLM APIs function as time-varying heterogeneous services rather than fixed model artifacts, with demand concentrated, supply-use mismatches, and task-specific routing yielding major cost and throughput gains.
Sparse prefix caching via dynamic programming for optimal checkpoint placement under overlap distributions improves the Pareto frontier for recurrent and hybrid LLM serving on shared-prefix data.
TrigReason matches large reasoning model accuracy on math and science benchmarks by delegating most steps to small models and intervening selectively on three triggers, cutting latency by 43.9% and cost by 73.3%.
Fleet adds a Chiplet-task level to GPU task models, enabling per-chiplet scheduling and cooperative cache reuse in persistent megakernels, yielding 1.3-1.5x lower LLM decode latency and up to 37% less HBM traffic on AMD MI350 hardware.
CodeComp uses Joern-extracted Code Property Graph priors for training-free structural KV cache compression, outperforming attention-only baselines on bug localization and code generation while matching full-context patch quality.
Stateful sessions with incremental KV cache and flash queries allow O(|q|) latency in streaming transformer inference, delivering up to 5.9x speedup over conventional engines while preserving full attention.
KV-RM regularizes KV-cache movement in static-graph LLM serving via block paging and merge-staged transport to improve throughput, tail latency, and memory use for variable-length decoding.
VisMMoE exploits visual-expert affinity via token pruning to achieve up to 2.68x faster VL-MoE inference on memory-constrained hardware while keeping accuracy competitive.
RaMP uses a hardware-derived performance region analysis and a four-parameter wave cost model to select optimal polymorphic kernel configurations for MoE inference from runtime expert histograms, delivering 1.22x kernel and 1.30x end-to-end speedups with 0.93% mean regret after brief profiling.
LLM agents avoid output stalling and reduce generation tokens by 48-72% via deferred template rendering guided by Output Generation Capacity and a Format-Cost Separation Theorem.
ProbeLogits performs single-pass logit reading inside the kernel to classify LLM agent actions as safe or dangerous, reaching 97-99% block rates on HarmBench and F1 parity or better than Llama Guard 3 at 2.5x lower latency.
Relax is a new RL training engine with omni-native design and async execution that delivers up to 2x speedups over baselines like veRL while converging to equivalent reward levels on Qwen3 models.
MEMENTO trains LLMs to segment reasoning into blocks, generate mementos as dense summaries, and reason forward using only mementos and KV states, cutting peak KV cache by ~2.5x while preserving benchmark accuracy.
Kernel-Smith combines evolutionary search with RL post-training to generate optimized GPU kernels, achieving SOTA speedups on KernelBench that beat Gemini-3.0-pro and Claude-4.6-opus on NVIDIA Triton and generalize to MetaX MACA.
Repeated sampling scales problem coverage log-linearly with sample count, improving SWE-bench Lite performance from 15.9% to 56% using 250 samples.
The paper delivers a unified executable benchmarking suite for tool-using agents that enforces a shared evidence-admission contract across web, code, and micro-task environments.
Shadow Mask Distillation enables KV cache compression in RL post-training of LLMs by mitigating amplified off-policy bias that defeats standard importance reweighting.
A server-side architecture with policy-aware ingestion and ABAC-based retrieval gating prevents cross-tenant data leakage in multitenant enterprise RAG and agent systems.
Training-free adaptive reuse of stable visual state in video VLMs reduces follow-up latency by 15-36x on Qwen2.5-VL while preserving correctness on VideoMME, with smaller first-query speedups via pruning.
citing papers explorer
-
VibeServe: Can AI Agents Build Bespoke LLM Serving Systems?
VibeServe demonstrates that AI agents can synthesize bespoke LLM serving systems end-to-end, remaining competitive with vLLM in standard settings while outperforming it in six non-standard scenarios involving unusual models, workloads, or hardware.
-
NCCLZ: Compression-Enabled GPU Collectives with Decoupled Quantization and Entropy Coding
NCCLZ decouples quantization and entropy coding across NCCL stack layers to enable overlapped compression, delivering up to 9.65x speedup over plain NCCL on scientific and training workloads.
-
Towards On-Policy Data Evolution for Visual-Native Multimodal Deep Search Agents
A new image-bank harness and closed-loop on-policy data evolution method raises multimodal agent performance on visual search benchmarks from 24.9% to 39.0% for an 8B model and from 30.6% to 41.5% for a 30B model.
-
Surviving Partial Rank Failures in Wide Expert-Parallel MoE Inference
EEP makes wide expert-parallel MoE serving survive single-rank failures with an 11s recovery pause, 8s reintegration pause, and throughput restored to 95% of pre-fault level within 52s while staying within 4.4% of a fixed-membership baseline in steady state.
-
Dooly: Configuration-Agnostic, Redundancy-Aware Profiling for LLM Inference Simulation
Dooly reduces LLM inference profiling costs by 56.4% via configuration-agnostic taint-based labeling and selective database reuse, delivering simulation accuracy within 5% MAPE for TTFT and 8% for TPOT across 12 models.
-
When Is the Same Model Not the Same Service? A Measurement Study of Hosted Open-Weight LLM APIs
Hosted open-weight LLM APIs function as time-varying heterogeneous services rather than fixed model artifacts, with demand concentrated, supply-use mismatches, and task-specific routing yielding major cost and throughput gains.
-
Sparse Prefix Caching for Hybrid and Recurrent LLM Serving
Sparse prefix caching via dynamic programming for optimal checkpoint placement under overlap distributions improves the Pareto frontier for recurrent and hybrid LLM serving on shared-prefix data.
-
TrigReason: Trigger-Based Collaboration between Small and Large Reasoning Models
TrigReason matches large reasoning model accuracy on math and science benchmarks by delegating most steps to small models and intervening selectively on three triggers, cutting latency by 43.9% and cost by 73.3%.
-
Fleet: Hierarchical Task-based Abstraction for Megakernels on Multi-Die GPUs
Fleet adds a Chiplet-task level to GPU task models, enabling per-chiplet scheduling and cooperative cache reuse in persistent megakernels, yielding 1.3-1.5x lower LLM decode latency and up to 37% less HBM traffic on AMD MI350 hardware.
-
CodeComp: Structural KV Cache Compression for Agentic Coding
CodeComp uses Joern-extracted Code Property Graph priors for training-free structural KV cache compression, outperforming attention-only baselines on bug localization and code generation while matching full-context patch quality.
-
Attention Once Is All You Need: Efficient Streaming Inference with Stateful Transformers
Stateful sessions with incremental KV cache and flash queries allow O(|q|) latency in streaming transformer inference, delivering up to 5.9x speedup over conventional engines while preserving full attention.
-
KV-RM: Regularizing KV-Cache Movement for Static-Graph LLM Serving
KV-RM regularizes KV-cache movement in static-graph LLM serving via block paging and merge-staged transport to improve throughput, tail latency, and memory use for variable-length decoding.
-
VisMMOE: Exploiting Visual-Expert Affinity for Efficient Visual-Language MoE Offloading
VisMMoE exploits visual-expert affinity via token pruning to achieve up to 2.68x faster VL-MoE inference on memory-constrained hardware while keeping accuracy competitive.
-
RaMP: Runtime-Aware Megakernel Polymorphism for Mixture-of-Experts
RaMP uses a hardware-derived performance region analysis and a four-parameter wave cost model to select optimal polymorphic kernel configurations for MoE inference from runtime expert histograms, delivering 1.22x kernel and 1.30x end-to-end speedups with 0.93% mean regret after brief profiling.
-
When Agents Go Quiet: Output Generation Capacity and Format-Cost Separation for LLM Document Synthesis
LLM agents avoid output stalling and reduce generation tokens by 48-72% via deferred template rendering guided by Output Generation Capacity and a Format-Cost Separation Theorem.
-
ProbeLogits: Kernel-Level LLM Inference Primitives for AI-Native Operating Systems
ProbeLogits performs single-pass logit reading inside the kernel to classify LLM agent actions as safe or dangerous, reaching 97-99% block rates on HarmBench and F1 parity or better than Llama Guard 3 at 2.5x lower latency.
-
Relax: An Asynchronous Reinforcement Learning Engine for Omni-Modal Post-Training at Scale
Relax is a new RL training engine with omni-native design and async execution that delivers up to 2x speedups over baselines like veRL while converging to equivalent reward levels on Qwen3 models.
-
MEMENTO: Teaching LLMs to Manage Their Own Context
MEMENTO trains LLMs to segment reasoning into blocks, generate mementos as dense summaries, and reason forward using only mementos and KV states, cutting peak KV cache by ~2.5x while preserving benchmark accuracy.
-
Kernel-Smith: A Unified Recipe for Evolutionary Kernel Optimization
Kernel-Smith combines evolutionary search with RL post-training to generate optimized GPU kernels, achieving SOTA speedups on KernelBench that beat Gemini-3.0-pro and Claude-4.6-opus on NVIDIA Triton and generalize to MetaX MACA.
-
Large Language Monkeys: Scaling Inference Compute with Repeated Sampling
Repeated sampling scales problem coverage log-linearly with sample count, improving SWE-bench Lite performance from 15.9% to 56% using 250 samples.
-
An Executable Benchmarking Suite for Tool-Using Agents
The paper delivers a unified executable benchmarking suite for tool-using agents that enforces a shared evidence-admission contract across web, code, and micro-task environments.
-
How to Compress KV Cache in RL Post-Training? Shadow Mask Distillation for Memory-Efficient Alignment
Shadow Mask Distillation enables KV cache compression in RL post-training of LLMs by mitigating amplified off-policy bias that defeats standard importance reweighting.
-
Securing the Agent: Vendor-Neutral, Multitenant Enterprise Retrieval and Tool Use
A server-side architecture with policy-aware ingestion and ABAC-based retrieval gating prevents cross-tenant data leakage in multitenant enterprise RAG and agent systems.
-
VLMaxxing through FrameMogging Training-Free Anti-Recomputation for Video Vision-Language Models
Training-free adaptive reuse of stable visual state in video VLMs reduces follow-up latency by 15-36x on Qwen2.5-VL while preserving correctness on VideoMME, with smaller first-query speedups via pruning.
-
SURGE: SuperBatch Unified Resource-efficient GPU Encoding for Heterogeneous Partitioned Data
SURGE achieves fixed-batch throughput for GPU embedding generation on 800M texts across 40k partitions using 12.6x less memory, 68x faster time-to-first-output, and fault tolerance via a streaming two-threshold policy with an analytical cost model accurate to 2%.
-
EdgeFM: Efficient Edge Inference for Vision-Language Models
EdgeFM is an agent-driven framework that strips non-essential features from VLMs and packages reusable optimized kernels, achieving up to 1.49x speedup over TensorRT-Edge-LLM on NVIDIA Orin while enabling first end-to-end deployment on Horizon Journey hardware.
-
LLM StructCore: Schema-Guided Reasoning Condensation and Deterministic Compilation
Two-stage Schema-Guided Reasoning with LLM condensation and deterministic compilation achieves macro-F1 of 0.63 on dyspnea CRF filling task and is language-agnostic.
-
HieraSparse: Hierarchical Semi-Structured Sparse KV Attention
HieraSparse delivers a hierarchical semi-structured sparse KV attention system that achieves 1.2x KV compression and 4.57x decode attention speedup versus prior unstructured sparsity methods at equivalent sparsity, plus up to 1.85x prefill speedup and 1.37x/1.77x speedups with magnitude pruning and
-
enclawed: A Configurable, Sector-Neutral Hardening Framework for Single-User AI Assistant Gateways
enclawed is a sector-neutral hardening framework for AI gateways providing signed modules, audit trails, peer attestation, and a 356-case test suite for regulated deployments.
-
Scalable Inference Architectures for Compound AI Systems: A Production Deployment Study
A deployed modular inference architecture for compound AI systems cut tail latency over 50%, boosted throughput up to 3.9x, and reduced costs 30-40% while handling multi-model agent workloads.
-
A Survey on Efficient Inference for Large Language Models
The paper surveys techniques to speed up and reduce the resource needs of LLM inference, organized by data-level, model-level, and system-level changes, with comparative experiments on representative methods.