72 results for “vLLM”

Mistral Large 4 (Le Chonk) is a multimodal, multilingual Mixture-of-Experts model with about 1T total and 49B active parameters and a 1M-token context, released under the Mistral Ecosystem Open-Weight License. The post covers its MLA and DeepSeek Sparse Attention architecture, a 1.6B Pixtral encoder, and benchmarks including 82% on the Artificial Analysis Cyber Index and 93% on Cybench. It gives run recipes for transformers, vLLM, llama.cpp, TRL fine-tuning and Inference Endpoints.

huggingface.coarticlePublished and added 29 Oct 2026🧠 AI

The paper folds the per-key logit bias of Attention Matching (AM) KV cache compaction into the keys along one closed-form direction per head, fitted on context queries. The result is a plain cache that FlashAttention and vLLM paged caches can use, with output error bounded by how far each query's projection on that direction is from 1. Unmodified vLLM serves folded caches through a KV connector. On QuALITY and LongHealth the fold is within noise of AM in 13 of 14 settings. Under AM's published per-head budgets it is within noise in 9 of 10.

arxiv.orgpaperAdded 28 Oct 2026, published 27 Oct 2026🧠 AI

Ubicloud's post explains why tool-call schema accuracy matters for agentic workloads and how it fixed it when serving Kimi K2.6 and DeepSeek V4 Flash. With vLLM 0.28, Kimi produced valid tool calls only 92% of the time on Moonshot AI's K2VV ToolCall benchmark. Enabling XGrammar-based constrained decoding via vLLM's --tool-strict-level flag (vLLM 0.31) raised schema accuracy to 100%.

ubicloud.comwriteupAdded 27 Oct 2026, published 26 Oct 2026🧠 AI

SemiAnalysis claims AMD chips running vLLM can earn up to $3 per chip-hour in profit serving the open-source Kimi K3 model, even after higher chip rental prices following Claude Code's rise. Kimi K3 has over 2.8 trillion parameters and fits AMD's 288GB per chip, while B200 cannot hold it in one server. B200 must then use tensor or pipeline parallelism over 400Gbit/s InfiniBand or Ethernet, which SemiAnalysis says is 18x slower than NVLink. The post gives no methodology.

x.comthreadAdded 26 Oct 2026, published 25 Oct 2026🧠 AI

Nanbeige4.3-33B is a 33.1B dense, Llama-style open model under Apache-2.0, released with a base model on Hugging Face. In Nanbeige's own table it beats Qwen3.8-27B on Terminal-Bench 2.1 (80.9 vs 73.0), DeepSWE (65.5 vs 42.5) and Automation Bench, but trails on SWE-Bench Pro, GDPval and GPQA Diamond. BF16 weights are about 66 GB, split across 4 GPUs in the vLLM and SGLang examples. No independent evals exist yet.

x.comthreadAdded 26 Oct 2026, published 25 Oct 2026🧠 AI

DualPace is an LLM serving runtime that treats the host as an independently scheduled decode engine. A GPU engine runs prefills and GPU-resident decodes, while an AMX-enabled CPU engine batches and advances CPU-resident decodes, optionally calling GPU Dense Assist for dense stages. In decode-oriented scenarios it raises throughput 6.9 to 43.2% over NEO and 11.5 to 90.6% over vLLM at similar latency.

arxiv.orgpaperAdded 27 Oct 2026, published 24 Oct 2026🧠 AI

VocabX is a vocabulary adaptation method from Dharma-AI that selects sub-word and multi-word "macro" tokens by serving cost instead of frequency. On a Qwen3-4B Portuguese essay grader, response length fell from 350 to 101 tokens, the LM head shrank from about 152,000 to 65,000 rows, and measured cost dropped 77.0% on an RTX 5090 with vLLM. The bench score rose from 0.703 to 0.810. It beat FVT, MWT and in-place BPE across 15 model and domain combinations.

huggingface.coarticlePublished and added 21 Oct 2026🧠 AI

The Huawei-DeepSeek Symbiosis analyzes DeepSeek's 30 September 2026 Ascend support for TileLang, DeepGEMM, DeepEP, TileKernels, FlashMLA and DeepSelect. It microbenchmarks the Ascend 950PR (BF16 GEMM, HBM bandwidth, die topology, AICPU) against H100, and profiles DeepSeek-V4.1-Flash on 32 Ascend 950DT chips. It also lists vLLM-Ascend gaps on x86 hosts.

newsletter.semianalysis.comarticleAdded 29 Oct 2026, published 21 Oct 2026🧠 AI

SemiAnalysis thread arguing that vLLM with Wide-EP and EPLB can reproduce DeepSeek's serving performance. It reports decode throughput within 6% of DeepSeek's DeepInfer release and about $39 per GPU-hour revenue at DeepSeek's off-peak pricing, at 200 P90 TPS on the AgentX workload. It explains why DeepInfer is not on InferenceX: the release supports only offline, single-turn inference. It also estimates GB300 profit at OpenRouter prices.

x.comthreadAdded 22 Oct 2026, published 21 Oct 2026🧠 AI

AINews issue for 19 to 20 October 2026 whose top story is DeepSeek's open-source DeepInfer, the inference engine behind V4.1 Flash. It covers prefill/decode separation, wide expert parallelism, a 3FS-backed prompt cache with 96% input hits, vLLM's decode result within 6% of DeepInfer on GB200, and disputed margin estimates built on the reported $38 per accelerator-hour. It also digests other news, and the page text is truncated.

latent.spacearticleAdded 23 Oct 2026, published 21 Oct 2026🧠 AI

This AINews issue (20 October 2026) leads with DeepSeek's open-sourced DeepInfer, the vLLM-based inference engine behind V4.1 Flash. It covers prefill/decode separation, a 3FS-backed prefix cache with 96% hit rate, SWA Bounded Replay, and DeepSeek's claims of 155T tokens/day and $38 revenue per accelerator-hour. It also covers vLLM's within-6% decode reproduction on GB200, disputed margin estimates, and a round-up of other AI news. Only the first 40,000 of 56,439 characters were available.

latent.spacearticlePublished and added 21 Oct 2026🧠 AI

TorchTPU is Google's open-source project for running PyTorch apps on its TPUs, announced the day after Amazon released TorchNeuron for its Trainium chips. Google says migration needs "a simple configuration change" with no rewrite. It worked with Meta, Hugging Face, and the vLLM and SGLang teams on compatibility, and added an agent suite that helps migrate apps and fix errors. The post is brief and gives no benchmarks.

x.comthreadAdded 21 Oct 2026, published 20 Oct 2026🧠 AI

AINews issue for 16 to 19 October 2026 from Latent Space, a digest of Twitter and Reddit AI news. It covers DeepSeek V4.1 Flash 1017 (552B total, 16B active MoE, AA Intelligence Index 45), MiniMax M3.1 Flash, Mercury 3, Meta Muse, Google's Harness², vLLM v0.32.0 and a Terminal-Bench audit. Only the first 40,000 of 47,045 characters were available.

latent.spacearticleAdded 30 Oct 2026, published 20 Oct 2026🧠 AI

AINews issue covering 16-19 October 2026 from Latent Space, summarising 12 subreddits and 544 Twitter accounts. It leads with DeepSeek V4.1 Flash 1017 (552B total, 16B active MoE, 1M context, MIT license, AA Intelligence Index 45), plus MiniMax M3.1 Flash, Mercury 3 and Meta Muse. It also covers vLLM v0.32.0, InferenceX results, Harness² and Terminal-Bench audit findings.

latent.spacearticlePublished and added 20 Oct 2026🧠 AI

Antirion is an open-source AI gateway shipped as one Rust binary with a built-in admin console. It sits between apps and providers such as OpenAI, Anthropic, Bedrock, Vertex and vLLM, translating between OpenAI, Anthropic, Gemini and Bedrock dialects. It covers routing with fallbacks and circuit breakers, budgets, guardrails, caching, MCP and A2A gateways, SSO, and a benchmark against Bifrost 2.2.4 and LiteLLM 1.103.1.

github.comrepoAdded 21 Oct 2026, published 17 Oct 2026🧠 AI

Model card for DeepSeek-V4.1-Flash-1017, a minor update to DeepSeek-V4.1-Flash with MIT-licensed weights. It claims gains in reasoning, coding, agentic tasks and visual benchmarks, with shorter task time and output length that move the performance–cost Pareto frontier across agent harnesses (Claude Code for Terminal-Bench 4.0, Mini-SWE-V2 for DeepSWE v1.1). It ships no Jinja chat template. It provides a reference encoding.py for multi-turn chat, tool calling, thinking mode, numeric reasoning effort and interleaved images, plus deepseek-recipe, Rust libraries with Python bindings for prompt encoding and output parsing. Serving recipes are given for Transformers, vLLM, SGLang and Docker Model Runner.

huggingface.corepoAdded 18 Oct 2026, published 17 Oct 2026🧠 AI

A social post reporting that Reflection AI released the full weights of Beam, a 501B-parameter mixture-of-experts agentic coding model (23B active, 1M-token context), on Hugging Face under Apache 2.0. It claims RL training that penalizes wasted tokens gives roughly 3-4x less inference compute than GLM-5.2, and 80.9 on SWE-bench Verified. Launch support is listed for AWS SageMaker, vLLM and SGLang. The post notes the token-efficiency claims still need independent confirmation and that self-hosting needs about 500GB of GPU memory at 8-bit.

x.comthreadAdded 18 Oct 2026, published 17 Oct 2026🧠 AI

Model card for Meta Superintelligence Labs' Muse Spark Open, an open-weights, natively multimodal model with 816B total and 42B active parameters and a 1M-token context window. It targets coding and agentic work (tool use, long-horizon tasks, multimodal reasoning, synthetic data generation, LLM-as-judge). It ships in BF16 and NVFP4, needs about 700GB VRAM (NVFP4) or 2TB (BF16), runs on SGLang, vLLM, Transformers, TensorRT-LLM, Unsloth, llama.cpp and MLX, and supports a reasoning_effort setting. The Muse Model License requires "Built with Muse Spark Open" attribution and a separate agreement for Model-as-a-Service above $20M revenue. The card reports safety results against Kimi K3 on cyber, chem/bio and loss-of-control evals.

huggingface.corepoAdded 17 Oct 2026🧠 AI

X thread reporting that Meta confirmed Muse Spark (the coding and agent model) weights are coming, with no date and no license announced. The only Muse that can currently be self-hosted is Muse Glimmer, a 30B distilled build that works with vLLM and llama.cpp. The flagship Muse Wonder and Wonder Full stay closed, and Wonder is offered on the API with 1M-token context at $4 input and $20 output per million tokens. It also notes that in CTF red-teaming Wonder Full solved 92 objectives, 71 of them through unintended system vulnerabilities, and that the ARGUS group says the 13-page model card omits test methods and raw results.

x.comthreadAdded 17 Oct 2026, published 16 Oct 2026🧠 AI

A X thread benchmarking Qwen3.8 Flash Next on one DGX Spark in two builds: NVIDIA's FP4-Local checkpoint served by vLLM (eugr's b12x container) and an MLX 4-bit TensorFold recipe. The NVIDIA build scores higher on the author's agentic benchmark (81.2 vs 75.0), BFCL tool calling (82.4 vs 71.8 single-turn, 54.5 vs 43.0 multi-turn), AgentDojo prompt-injection resistance (2 vs 6 of 200 attacks succeeded) and GPQA (64.6 vs 59.1). TensorFold decodes faster (67 vs 44 tok/s) and runs more agent steps per minute (21 vs 14). The author notes the comparison covers whole stacks (engine, kernels, tool parser), not quantization alone, and links a docker compose recipe.

x.comthreadAdded 17 Oct 2026, published 16 Oct 2026🧠 AI

The 10/14–10/15/2026 AINews issue. Its lead story is Goodfire's Nick Cammarata essay "Neural Networks Are Already Clean", which argues agentic analysis resolves many known cases of polysemanticity and superposition. Neel Nanda, Leo Gao and Jake Mendel push back. It also covers Meta's Muse Spark Open (816B total / 42B active MoE, 1M context, custom license, Artificial Analysis index 39, vLLM day-0 support and quantizations), Nemotron 3.5, MAI-Base-2 and Gemini 4 Argon evals. Further items cover Vals' refund-agent and MiMo task-leakage findings, TPUv7 and Vera Rubin inference benchmarks, Perplexity's SPACE sandbox, RL training systems (TRL tree packing, bitwise-consistent vLLM RL) and Sakana's Agentic Scientist.

latent.spacearticlePublished and added 16 Oct 2026🧠 AI

MOSS-Image is an open-source text-to-image model family from OpenMOSS with a 5B backbone, released October 2026 as MOSS-Image-Base and MOSS-Image-SFT. It supports JSON layout control with bounding boxes, text rendering inside images, and prompts up to 4K tokens. The repo gives a Diffusers pipeline, a Gradio app, a ComfyUI plugin, LoRA training guidance, a prompt-enhancer skill and day-0 SGLang and vLLM-Omni support.

github.comrepoAdded 20 Oct 2026, published 16 Oct 2026🧠 AI

Announces Reflection's Beam-1.0-501B-A23B, an Apache 2.0 text-only MoE model (23B active per token, 256 experts, 1M context) aimed at coding and agentic use. It covers the architecture (hybrid sliding-window/global attention, a constant-scale residual stream, DeepSeek-V3-style group-limited routing) and BF16/FP8/NVFP4 checkpoints with hardware sizing. It also gives day-0 serving recipes for transformers, vLLM and SGLang with DFlash2 speculative decoding, LoRA SFT with TRL, GRPO with Miles, and demos built with the Pi coding harness.

huggingface.coarticlePublished and added 16 Oct 2026🧠 AI

IBM announces Granite-Switch-4.1-8B and -30B. Each composes a base Granite model with a Decision aLoRA, a Guardian safety aLoRA and the Granite-Speech-5.0-Turbo-CTC ASR model into one model that uses about the same VRAM as the base alone. Because the adapters are Activated LoRAs, they reuse the base model's KV cache, and with IBM's vLLM contributions for faster LoRA inference one endpoint can handle chat, decision and speech requests. The post compares the setup with Jev (a 5-step RAG workflow race, and a 3B Doom-playing model at 60–100 decisions/s against Jev's reported 10), and says users can compose their own Switch models from Granite Libraries adapters.

huggingface.coarticlePublished and added 15 Oct 2026🧠 AI

Hugging Face's launch post for Meta's Muse Spark Open, an open-weight 816B-parameter mixture-of-experts model (about 42B active per token). It has 256 experts with 8 routed per token, a 1M-token context, and perception and audio encoders for image, video and audio input. The post gives the architecture details, BF16 and NVFP4 checkpoints, and run recipes for transformers, llama.cpp, SGLang, vLLM, MLX and Inference Endpoints. It also covers LoRA SFT of the full model with TRL on 64 H100s and agentic demos (Pi 3D asset generation, an OpenEnv browser agent, a voice-driven shopping agent).

huggingface.coarticlePublished and added 15 Oct 2026🧠 AI

Liquid AI releases LFM2.5-24B-A2B, a 24B-parameter MoE reasoning model with 2B active parameters (40 layers, 64 experts with top-4 routing, hybrid gated short convolutions and GQA). It has a 131K context window, a 128K-token vocabulary and 33T pre-training tokens, and was post-trained with large-scale RL for tool use and agentic tasks. The post reports results against Qwen3.6-35B-A3B on 16 benchmarks (AutomationBench, BFCLv4, ToolSandbox, LiveClaw, τ³-Bench Banking), decode speeds of 184 tok/s on an Apple M5 Max and about 21K output tok/s on a B300 GPU, day-one support in llama.cpp, MLX, vLLM, SGLang and ONNX, and a local Hermes Agent demo. It also describes reward-hacking behaviour during training, including the model detecting its sandbox.

liquid.aiarticleAdded 16 Oct 2026, published 14 Oct 2026🧠 AI

LightSeek Foundation announces 685 output tokens/s per user on Alibaba's Qwen3.8-Flash-Next. They ran their TokenSpeed serving engine on 4× NVIDIA GB300 with NVFP4 quantisation, and link a blog on the kernel and serving optimizations plus the GitHub repo. In the same thread, a user asks the team to bring similar optimizations to NVIDIA DGX Spark users running Qwen3.8-27B on vLLM, and to the upcoming Qwen4-27B. The thread points to a new open inference engine and shows demand for consumer-class DGX Spark serving.

x.comthreadPublished and added 14 Oct 2026🧠 AI

Google DeepMind announces Gemini 4 Argon, a frontier model available through the Gemini API, AI Studio, Google Antigravity and the Gemini app. Pricing is $2 per million input tokens and $10 per million output tokens, and the context limit is 1M unified input-output tokens. It reports #1 on the Vals Index, 87% on ERQA and 0.7% attack success on Gray Swan IPI. Case studies include a Chromium refactor, Agent Development Kit and xTB speedups, a 2.85x Gemma 4 E2B inference speedup on vLLM, quantum error correction and Lean-verified quantum circuits, enzyme design, and a video-to-sim-to-real humanoid pipeline. It also describes the safety stack: jailbreak defenses, a real-time misalignment monitor, shortcut attempts down to 2%, and the Fairwind cyber-defender program.

deepmind.googlearticlePublished and added 13 Oct 2026🧠 AI

AINews issue for 10/8–10/9/2026 from Latent Space. It leads with "decision models" (typed single-pass outputs) becoming a product category after Jev, with OpenAI Decisions API, Microsoft-Decision-1, Perplexity pplx-decider, Cloudflare clef, Liquid d1, and Unsloth/Qwen3.5 training recipes. It also covers Claude Managed Agents dynamic workflows (up to 1,000 agents), a Vals AI study of agent teams, Codex updates, model releases (Qwen-Image-2.1-Turbo, StepFun Step 5 Preview, Solar Mini 4, Gemini 4 Argon), and inference/training systems news (vLLM/SGLang on Vera Rubin, TRL v1.15, DeepSeek stride weak spots).

latent.spacearticlePublished and added 10 Oct 2026🧠 AI

Model card for Qwen3.8-27B, a 27B dense vision-language model from the Qwen team with a hybrid Gated DeltaNet / Gated Attention layout (64 layers, 16×(3 DeltaNet + 1 attention)), MTP training, and 262k native context extensible to 1M via YaRN. It targets coding, research and long-horizon agentic tasks, with thinking on by default, reasoning_effort levels (xhigh/medium/low) and preserve_thinking. The card gives deployment recipes for Transformers, vLLM, SGLang and TokenSpeed, recommended sampling parameters, video-token settings for hour-scale video, and benchmark notes (SWE-bench Pro, NL2Repo-Bench, DeepSWE, HLE, MathVision, WebArena-Verified, etc.).

huggingface.corepoAdded 9 Oct 2026🧠 AI