Release24d ago
OpenBMB open-sources MiniCPM-V 4.7: a 35B-A3B MoE multimodal model that runs on-device at 13+ tokens/s on iPhone 17 Pro Max
On 2026-10-12 OpenBMB released MiniCPM-V 4.7, an Apache-2.0 open multimodal family with a 35B-A3B variant (the series' first MoE) and a 1B variant. Its Lookahead Expert Scheduling (LES) prefetches experts across layers and, combined with layer-wise mixed-precision quantization and DSpark speculative sampling, lifts SSD-offloaded decoding on an iPad Pro M4 (16GB) from 2.97 to 17.39 tokens/s (5.86x), per OpenBMB. Weights are on Hugging Face, ModelScope and Ollama, with vLLM, SGLang, llama.cpp and iOS/Android/HarmonyOS support.
Tooling10d ago
Google's Gemma team says Gemini 4 Argon wrote vLLM optimizations that speed up gemma-4-E2B-it 2.85x on an A100
On Oct 26, Google's official @googlegemma account said a set of vLLM optimizations built with Gemini 4 Argon lifted gemma-4-E2B-it throughput on a single A100 from about 37,900 to 108,000+ tokens/sec, roughly 2.85x, with 100% token parity. The PRs were filed by Luciano Martins of Google DeepMind, and most are still under review.
Tooling16d ago
vLLM and Ant Group open-source Humming, one quantized-inference kernel library for many formats and every NVIDIA GPU since Turing, lifting DeepSeek-V4.1-Flash end-to-end throughput by 14-43%
On Oct 21 the vLLM community and Ant Group open-sourced Humming (Apache 2.0), a quantized-inference library that JIT-compiles GEMM kernels on demand. It supports every NVIDIA architecture from Turing on plus T-Head's Zhenwu 810E, and covers weight and activation combinations from INT1-INT8 and FP4-FP8 for dense and MoE workloads. In vLLM's microbenchmarks it runs 2-5x faster than vLLM Marlin, FlashInfer or DeepGEMM in some cases; on 8 H20 GPUs, switching DeepSeek-V4.1-Flash MoE from Marlin W4A16 to Humming W4A8 raised throughput 14-43%. vLLM 0.29.0 ships it built in.
Discussion & Comments
0Sign in to join the discussion
Connect with AI developers to exchange benchmark insights.