⚡ Each donation funds the next large quant.

I host free GGUF or MoE quants as independent research.
Local hardware: Mechrevo Kuangshi GM7AG0M — RTX 3060 Laptop 6GB GDDR6, 64GB DDR5, i7-12700H (14C/20T, 4.7GHz), Windows 11, Samsung 990 Pro.
Good for imatrix and 0.6–35B-class work in RAM. 9B+ and searches need rented H200/Blackwell, typically $100 per quant.

🎉 Boosty🦄  |  ☕ Buy Me a Coffee🦄  |  ⭐ DonationAlerts🦄

💚 Thanks to Hugging Face for extra storage.🦄

---

NOESIS / AMAImedia

Released as part of the NOESIS Professional Multilingual Dubbing Automation Platform (framework: DHCF-FNO — Deterministic Hybrid Control Framework for Frozen Neural Operators)

• Founder: Ilia Bolotnikov
• Organization: AMAImedia.com
• X (Twitter): @AMAImediacom
• LinkedIn: Ilia Bolotnikov
• Telegram: @djbionicl
• Release date: 2026-09-27

AMAImedia

• Original model: XiaomiMiMo/MiMo-V2.6-Flash-MOPD

• GGUF (source of this mirror): AesSedai/MiMo-V2.6-Flash-MOPD-GGUF

• Quantizer: ed's bpw-size PR (llama.cpp #15550)

---

| 🤗 HuggingFace  | 📰 Blog  | 🎨 Xiaomi MiMo API Platform  | 🗨️ Xiaomi MiMo Studio  | 💻 Xiaomi MiMo Desktop  |
Community
WeChat Group  |  Discord  |  Telegram  |  Reddit

Upstream model card. The block below is reproduced from the official XiaomiMiMo/MiMo-V2.6-Flash-RL card. It describes the MiMo-V2.6 series base model — architecture, training recipe and benchmark numbers for the BF16/FP8 checkpoints. It is not a measurement of this repository's BPW2.5 GGUF. For the quant recipe, verified metrics and file list of this mirror, see below.

MiMo-V2.6-Flash-RL

Scaling Reinforcement Learning Toward Self-Improvement

Technical Report

1. Introduction

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.
• You Only RL Once: One mixed RL run across coding, general agents, visual, and cybersecurity — not separate per-domain runs. Tasks and multiple harnesses are mixed in the same batch so capabilities reinforce each other and strategies transfer to harnesses never seen in training.
• Scaling RL Compute: Fully asynchronous Group Relative Policy Optimization (GRPO) on very large batches — 1,568 prompts × 16 rollouts per step, billions of tokens per update.
• Groupwise Agentic Grading (Self-Improvement Loop): Binary pass/fail cannot rank passing solutions, so the reward signal itself is scaled. An agentic grader compares rollouts within each group: Groupwise Reward Synthesis (GRS) builds task-specific rubrics offline from contrasting rollouts and fuses rubric quality with test outcomes; Groupwise Advantage Redistribution (GAR) ranks passing trajectories online and moves advantage toward higher-quality solutions. Judged against the policy’s own samples, this closes a self-improvement loop and steers toward shorter paths and fewer tokens per task.
• Aligned RL: Cold start from self-correction — the model reflects on and rewrites its own misaligned turns into grounded next steps. Throughout RL, environment hardening, adversarial screening, and verifier cross-checks keep the loop honest against reward hacking.
• Multi-Prefix Multi-Teacher On-Policy Distillation (MOPD2): After mixed RL, MOPD2 combines autonomous student rollouts with prefix-conditioned single-turn rollouts (Teacher-Prefix and SFT-Prefix), reusing histories from teacher trajectories and SFT demonstrations so decision points train without regenerating preceding turns — extending capabilities to hard-to-verify tasks.

Model Summary

• Architecture: Sparse MoE (Mixture of Experts), 309B total / 15B activated parameters
• Context Length: 1M tokens
• Modalities: Text, Image, Video, Audio
• Vision Encoder: 681M-param MiMo ViT (28 layers: 24 SWA + 4 Full)
• Audio Encoder: 308M AudioTokenizer + 127M audio patch encoder
• Multi-Token Prediction (MTP): 5-layer speculative decoder

Figure 1. MiMo-V2.6 architecture.

3. Evaluation Results

4. Model Architecture

LLM Backbone

The first Transformer block uses global attention with a dense FFN. Remaining blocks interleave local SWA and GA; both use sparse MoE FFNs without shared experts.

Vision Encoder (MiMo ViT)

Audio Encoders

AudioTokenizer encoder: 24 layers (12 SWA / 12 GA), hidden 1024, 20 RVQ codebooks, 308M parameters. Audio patch encoder: 6 layers, 127M parameters; four frames per patch (25 Hz → 6.25 Hz).

Speculative Decoder

5-layer SWA MTP drafter (DFlash-style). Predicts 7 subsequent tokens per forward pass for parallel verification.

Citation

@misc{mimo2026v26flash,
  title={MiMo-V2.6-Flash-RL},
  author={{Xiaomi MiMo Team}},
  year={2026},
  howpublished={\url{https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL}},
}

Contact

For questions or feedback, reach us at mimo@xiaomi.com or join our community:

• WeChat Group
• Discord
• Telegram
• Reddit

---

MiMo-V2.6-Flash-MOPD — BPW2.5 GGUF

This is a mirror of the upstream BPW2.5 quant by. The quantization was not re-run by AMAImedia. The three shards were copied server-side and their LFS OIDs and byte sizes were verified identical to the source repository. No re-quantization, no re-calibration, no re-packaging of the tensor payload was performed here.

Upstream license and terms apply. Weights derive from (MIT).

What was mirrored

Only the BPW2.5 tier plus the multimodal projector and the KLD/PPL analysis data. Other tiers (BPW2.0, BPW3.0, BPW3.5, MXFP4) remain upstream-only.

Total main payload 96.74 GB. Download all three shards into the same directory and keep the projector alongside them.

Upstream quant recipe

The idea: the FFN tensors dominate total size, so quality is preserved by keeping the non-FFN tensors at high precision and quantizing FFN GATE + FFN UP + FFN DOWN down to the BPW target. The rest of the model stays at Q8_0 / Q6_K. BPW quants were produced with ed's bpw-size PR.

For BPW2.5 the non-FFN mixture is Q6_K, and the FFN tensors are quantized down to reach 2.50 effective BPW.

Upstream measured quality

Measured by the upstream quantizer; reproduced here unchanged.

Reference base PPL ≈ 5.143516 (derived from the MXFP4 row, which is the full-quality tier).

Per-quant detail: BPW2.5 notes, full dataset.

Usage

Requires a llama.cpp build with BPW quant type support — upstream states the BPW quants require llama.cpp PR #15550. The projector also requires MOPD multimodal support in the same runtime.

Note the model is very large. A 2.50 BPW quant of this MoE is roughly 90 GiB of weights, so plan host RAM/VRAM accordingly.

License and attribution

Upstream did not publish a separate license tag for the GGUF folder, so the upstream model card terms apply. Please cite the upstream quantizer rather than this mirror.