DeepSeek-V4-Flash-DSpark

Model · Text generation

MiMo-V2.6-Flash-RL

Xiaomi MiMo

Scaling Reinforcement Learning Toward Self-Improvement MiMo-V2.6-Flash-RL is the efficiency-balanced checkpoint of the MiMo-V2.6 series. The series is built to scale reinforcement learning toward self-improvement — scaling RL compute, environment diversity, and grader compute together, so the model keeps expanding its capability frontier through exploration and feedback. Key features include: - Native Omnimodal + Long Horizon: Text, image, video, and audio in one model; 1M tokens for long repositories, tool traces, and multi-session agent runs. - Multi-Prefix Multi-Teacher On-Policy Distillation (MOPD2): After mixed RL, MOPD2 combines autonomous student rollouts with prefix-conditioned…

Model · Text generation

MiMo-V2.6-Flash-RL

Zachary Howard

Scaling Reinforcement Learning Toward Self-Improvement MiMo-V2.6-Flash-RL is the efficiency-balanced checkpoint of the MiMo-V2.6 series. The series is built to scale reinforcement learning toward self-improvement — scaling RL compute, environment diversity, and grader compute together, so the model keeps expanding its capability frontier through exploration and feedback. Key features include: - Native Omnimodal + Long Horizon: Text, image, video, and audio in one model; 1M tokens for long repositories, tool traces, and multi-session agent runs. - Multi-Prefix Multi-Teacher On-Policy Distillation (MOPD2): After mixed RL, MOPD2 combines autonomous student rollouts with prefix-conditioned…

Model · Text generation

Hy3-Razor-154B-A18B-E96of192

Mingyang Song

Hy3 with half of its routed experts removed by RAZOR, a training-free expert pruning method. Every MoE layer keeps 96 of its original 192 routed experts. No gradient updates or recovery training were applied: the retained weights are the base model's own weights. Pruning touches only the routed expert pool. Attention, shared experts, the embedding and the LM head are untouched, so the compute per token drops only by the share of expert FLOPs that the removed experts would have contributed. The other budget is Requires a Transformers build containing the native hyv3 implementation. Weights are bfloat16. RAZOR asks whether the surviving computation can replace an expert's function, rather…

Goal: oQ4e speed and memory footprint, oQ5e-or-better quality reaching toward oQ6e. Runs at ~73–75 GB resident with ~82→66 tok/s decode on a 128 GB Apple Silicon host — the speed/RAM economics of a 4-bit build, at an accuracy tier equal to or above uniform oQ5e. How: a mixed-precision MLX build of Qwen3.8-Flash-Next, assembled by component-wise byte-copy from Jundot's native oQ quantizations — zero re-quantization in the entire package. Every tensor is byte-identical to its source, so quality is traceable, not re-derived. Uniform-bit quantization pays accuracy where it isn't needed (middle layers) and loses it where it hurts (first/last layers, n-gram table, attention). A Data-Free…

Full GLM-5.3 (753B, glmmoedsa) quantized to EXL3 with a per-expert mixed bit width averaging 2.75 bpw on the routed experts, sized to serve on four NVIDIA DGX Sparks (GB10, 128 GB unified memory each) with room for a 200K-token KV pool (or 1M tokens with decode-context-parallel 4). Unmodified weights of zai-org/GLM-5.3 (no abliteration, no fine-tuning), quantized from the BF16 checkpoint. For the earlier uniform-width quant see The fastest way to serve this checkpoint is TensorFold's native glmmoedsa engine — tensor parallel 4, one rank per Spark, no vLLM in the serving path; a drafted reply equals a serial one. Recipe, one-shot launcher, benches and an AGENTS.md needs no extra draft model…

Model · Text generation

Qwen3.8-Flash-Next-P48NVFP4-MoESQ

IST Austria Distributed Algorithms and Systems Lab

A W4A4 + paired-4:8 sparse compressed checkpoint of Qwen/Qwen3.8-Flash-Next, produced with MoESQ. The routed MoE expert weights are NVFP4 with paired-4:8 structured sparsity and are stored sparse: only the kept values plus a small mask are on disk. The target is NVIDIA Blackwell (SM100) sparse tensor cores. expert, 48 layers: 36 linear-attention and 12 sparse-attention, hyper-connections, per-layer n-gram embedding) weight including the sparsity mask and scales (2 bits per weight for the kept values). TP2). The per-layer n-gram embedding tables (95.4 GiB, BF16, unchanged) are lookup tables served from CPU memory, as in the base model. paired48nvfp4 MoE backend Links This checkpoint does not…