> [!Note] > This is an improved version of Kimi-VL-A3B-Thinking. Please consider using this updated model instead of the previous version. > [!Note] > Please visit our tech blog for recommended inference recipe of this model: Kimi-VL-A3B-Thinking-2506: A Quick Navigation Comparison with efficient models and two previous versions of Kimi-VL (*Results of GPT-4o is for reference here, and shown in italics ):
Recently, the Muon optimizer has demonstrated strong results in training small-scale language models, but the scalability to larger models has not been proven. We identify two crucial techniques for scaling up Muon: - **Weight Decay**: Critical for scaling to larger models - **Consistent RMS Updates**: Enforcing a consistent root mean square on model updates These techniques allow Muon to work out-of-the-box on large-scale training without the need of hyper-parameter tuning. Scaling law experiments indicate that Muon is $\sim2\times$ more sample efficient than Adam with compute optimal training.
Recently, the Muon optimizer has demonstrated strong results in training small-scale language models, but the scalability to larger models has not been proven. We identify two crucial techniques for scaling up Muon: - **Weight Decay**: Critical for scaling to larger models - **Consistent RMS Updates**: Enforcing a consistent root mean square on model updates These techniques allow Muon to work out-of-the-box on large-scale training without the need of hyper-parameter tuning. Scaling law experiments indicate that Muon is $\sim2\times$ more sample efficient than Adam with compute optimal training.
**MoonViT** is a Native-resolution Vision Encoder, which is initialized from and continually pre-trained on **SigLIP-SO-400M**. To facilitate the standalone use of MoonViT, we have separated the implementation and weights of MoonViT from moonshotai/Kimi-VL-A3B-Instruct. If you are interested in the training process of MoonViT, you are welcome to read Paper Kimi-VL Technical Report.
Version 0.5 (git tag `v0.5`). See the changelog. Weight set ID: `2011c0cedde962d19fc43e74c1d409279e398d40d27a6fdb4bcb7b49ea2ea6fb` (see Identifying the weights). Also mirrored on ModelScope: kindlingai/glm-5.3-exlr8-k2-k3-k4. > **These weights cannot be used without E4M3 rounding.** A loader must round every value the trellis > codebook produces to FP8 E4M3 (round to nearest even, saturating) before applying the side vectors. The > encoder searched only over E4M3-rounded values, so the rounding is part of the format: a stock EXL3 decoder, > or any decoder that skips the rounding, reads measurably worse weights from the same bytes.
A quantization of **GLM-5.3-Flash** in the **exlr8 v2** format, built for **NVIDIA GB10**-class unified-memory devices (DGX Spark, ASUS GX10) served on one to four boxes. Every routed and shared expert is stored at three trellis widths, K2, K3 and K4, and the serving width is chosen at load time. > **These weights cannot be used without E4M3 rounding.** A loader must round every value the trellis codebook > produces to FP8 E4M3 (round to nearest even, saturating) before applying the side vectors. The encoder searched > only over E4M3-rounded values, so the rounding is part of the format: a stock EXL3 decoder, or any decoder that > skips the rounding, reads measurably worse weights from the same bytes. GLM-5.3-Flash is a MoE model: 45 layers (0-2 dense, 3-44 MoE), 288 routed experts with top-8 routing plus one shared expert per MoE layer, an MTP layer 45 and a vision tower. This repository quantizes the routed and shared experts of layers 3-45 and the dense weights of every layer, and carries the vision tower unchanged in BF16.
> **Tip**: Use `--no-think` or `/no_think` mode for faster responses and quality. The thinking tokens consume output budget at ~3.4 tok/s, so disabling thinking gives you more usable output** **This model does not refuse.** It provides complete, detailed, working answers on every tested category. Fine-tuned on the FableForge Mix A dataset (47,824 examples) of agent traces, shell commands, code generation, and multi-step reasoning tasks. 98.3% of the 2.8M formatted examples remain untapped for future training.
A 1.5B model scoring **96%** — punching far above its weight. 30-prompt evaluation (domain knowledge, uncensored domain + general, reasoning, tool use), 0–5 scoring, run on an NVIDIA A40 via Ollama.
A 1.5B model scoring **93%** — punching far above its weight. 30-prompt evaluation (domain knowledge, uncensored domain + general, reasoning, tool use), 0–5 scoring, run on an NVIDIA A40 via Ollama.
A 1.5B model scoring **93%** — punching far above its weight. 30-prompt evaluation (domain knowledge, uncensored domain + general, reasoning, tool use), 0–5 scoring, run on an NVIDIA A40 via Ollama.
