Qwen3.5-35B-A3B

Headlines · 3

1. 5d
   WorkRover trains long-horizon cowork agents via recursive environment evolution

   An arXiv paper introduces WorkRover, a framework for training long-horizon cowork agents through recursive environment evolution: its harness-driven synthesis has construction agents jointly build semantically aligned work contexts, executable tasks and evidence-grounded verifiers, adapting their actions to workspace state and execution feedback instead of following a fixed pipeline. To support environment evolution the authors propose Factorial Compatibility Replay (FCR), which uses rollout feedback to evaluate combinations of component revisions and selects the smallest validated repair, plus Verifier-to-Turn Credit Assignment (VTCA), which augments trajectory-level GRPO advantages with deterministic workspace-progress signals. Post-training Qwen3.5-35B-A3B with WorkRover yields WorkRover-35B-A3B, which the authors report exceeds the base model by 26.27 percentage points on JobBench, 29.32 points on OfficeQA Pro and 413.25 Elo points on GDPval v2.

   Hugging Face · Papers · 🔥 0
2. 6d
   FramePack-U packs hour-long video into a fixed token budget at constant memory

   The paper "Elephant in the Fridge: Constant-Memory Frame Packing for Long Video Understanding" introduces FramePack-U, a training-free frame-packing framework that compresses arbitrarily long videos into a fixed-size visual representation at constant memory cost, requiring no fine-tuning or architectural change and plugging directly into off-the-shelf vision-language models (VLMs) at inference time. It scores each frame by semantic similarity to the text prompt, ranks frames in a diversity-driven manner, and packs the selected frames at varying resolutions into a fixed-length context window so the memory budget goes to frames that are both instruction-relevant and visually diverse. The authors report that on LongVideoBench, MLVU and Video-MME, FramePack-U consistently beats full-resolution baselines and prior token-reduction and keyframe-selection methods, including gains of +12.9%, +10.6% and +9.7% over the official model on Qwen3.5-35B-A3B with a 128-frame budget.

   Hugging Face · Papers · 🔥 0
3. Oct 9
   SGLang and Miles tuned for NVIDIA Vera Rubin, serving Kimi K3 NVFP4

   The SGLang & Miles team says it received early access to two NVIDIA Vera Rubin nodes (8 GPUs) and brought both SGLang inference and Miles RL training up on the platform. For serving the Kimi K3 NVFP4 checkpoint, it retuned attention, MoE/communication and KDA kernels, reporting a 16% faster deeper MLA pipeline at batch 16 and 128k context, a 20% faster split-KV reduction, a 5.9% end-to-end gain from fusing the MoE tail, and 276 fewer kernel launches per decode step. On the training side, Miles uses SGLang for rollout and Megatron for training, and ran end-to-end RL out of the box on a single 4-GPU tray with Qwen3-30B-A3B and DeepSeek-V4-Flash, plus agentic RL with 64 sandboxes on the Arm-based Vera CPU on SWE-bench Verified. The team also cites Cognition as reporting up to a 4.8x increase in total token throughput on Vera Rubin NVL72 versus GB200 NVL72 using its own private SGLang fork.

   LMSYS Blog · 🔥 0

Experience and discussion from the community

Share my Qwen3.5-35B-A3B experienceAsk about Qwen3.5-35B-A3B