Saga Ishtardottir writes that CyberTiel is an abliterated, quantized MoE model designed for high-speed agentic coding and offensive security tasks. Based on the Ornith-1.5-35B-A3B architecture, this version features a grafted multi-token prediction (MTP) head to enable speculative decoding via llama.cpp and uses a custom cyber-weighted importance matrix to maintain performance in low-bit quantizations. The model is optimized for real-world programming problems, showing significant improvements on benchmarks like SWE-bench-Live compared to its censored counterparts.
- Includes an abliterated (uncensored) base that allows the model to perform offensive security work without refusal.
- Features a grafted MTP head at block 40 specifically designed for speculative decoding speedups in llama.cpp.
- Uses Unsloth Dynamic quantization with a custom imatrix calibrated on specialized cybersecurity and code datasets.
- Capable of vision tasks via the mmproj-BF16.gguf projector from Ornith's original model.
- Includes an abliterated (uncensored) base that allows the model to perform offensive security work without refusal.
- Features a grafted MTP head at block 40 specifically designed for speculative decoding speedups in llama.cpp.
- Uses Unsloth Dynamic quantization with a custom imatrix calibrated on specialized cybersecurity and code datasets.
- Capable of vision tasks via the mmproj-BF16.gguf projector from Ornith's original model.
This repository provides quantized GGUF versions of the Tiel coder model, which is a re-quantized and dynamically improved version of Ornith-1.5 35B designed specifically for agentic coding tasks. It excels at multi-turn conversations and solving codebase issues with performance comparable to Opus 4.6 medium in certain benchmarks.
- Optimized via an importance matrix (imatrix) using a calibration corpus heavily weighted toward code.
- Inherits the Ornith-1.5 vision tower, allowing for image-to-text tasks like analyzing screenshots of failing tests.
- Uses the Sharp chat template to produce more concise answers compared to its base model.
- Designed with a Mixture of Experts (MoE) architecture that enables fast inference and small KV cache footprint in RAM.
- Optimized via an importance matrix (imatrix) using a calibration corpus heavily weighted toward code.
- Inherits the Ornith-1.5 vision tower, allowing for image-to-text tasks like analyzing screenshots of failing tests.
- Uses the Sharp chat template to produce more concise answers compared to its base model.
- Designed with a Mixture of Experts (MoE) architecture that enables fast inference and small KV cache footprint in RAM.
Dmitry Soldatkin, Andrew Smith, and Vinay Arora write about deploying the open-weight Qwen3.8-2.4T-A95B model on Amazon SageMaker HyperPod using vLLM to support demanding agentic and reasoning workloads. The article provides a technical walkthrough for hosting this massive 2.4 trillion parameter mixture-of-experts (MoE) model on NVIDIA B300 Blackwell Ultra GPUs, covering infrastructure sizing with NVFP4 quantization, configuration of features like Multi-Token Prediction (MTP), and performance optimization through expert parallelism and prefix caching.
- The model uses a hybrid architecture combining Gated DeltaNet layers for linear attention and Gated Attention layers for full quadratic attention to manage long context windows up to 1 million tokens.
- NVFP4 quantization reduces the model's weight footprint to ~1.2 TB, allowing it to fit on a single node with 8× NVIDIA B300 GPUs.
- Enabling Multi-Token Prediction (MTP) speculative decoding can reduce Time-To-First-Token (TTFT) by nearly 60%.
- Deployment is managed via the SageMaker HyperPod Inference Operator using Kubernetes (EKS) for automated lifecycle management and resilience.
- The model uses a hybrid architecture combining Gated DeltaNet layers for linear attention and Gated Attention layers for full quadratic attention to manage long context windows up to 1 million tokens.
- NVFP4 quantization reduces the model's weight footprint to ~1.2 TB, allowing it to fit on a single node with 8× NVIDIA B300 GPUs.
- Enabling Multi-Token Prediction (MTP) speculative decoding can reduce Time-To-First-Token (TTFT) by nearly 60%.
- Deployment is managed via the SageMaker HyperPod Inference Operator using Kubernetes (EKS) for automated lifecycle management and resilience.
Moonshot's newly released Kimi K3 is a 2.8 trillion parameter model that marks their most capable release to date. This article explores its performance benchmarks, pricing structures compared to Anthropic’s Sonnet series, and how it handles complex tasks like SVG generation. Through the lens of the pelican benchmark, the author examines the relationship between reasoning token consumption, cost, and a model's spatial awareness.
Main points:
- K3 is described as an open 3T-class model with high performance on long-horizon knowledge work.
- The pricing structure for K3 represents a significant increase over previous Moonshot models.
- Testing the pelican prompt reveals high reasoning token usage and substantial costs per task.
- While useful for checking spatial awareness, simple benchmarks fail to test agentic tool calling.
Main points:
- K3 is described as an open 3T-class model with high performance on long-horizon knowledge work.
- The pricing structure for K3 represents a significant increase over previous Moonshot models.
- Testing the pelican prompt reveals high reasoning token usage and substantial costs per task.
- While useful for checking spatial awareness, simple benchmarks fail to test agentic tool calling.
Google's release of Gemma 4 marks a major turning point for open-source AI, offering a versatile family of multimodal models under a permissive Apache 2.0 license. Built using Gemini 3 technology, these models demonstrate massive leaps in math and coding performance, rivaling much larger proprietary systems while remaining efficient enough to run on local hardware ranging from smartphones to high-end GPUs. This release positions Google as a formidable competitor in the open-weights ecosystem, prioritizing user ownership and deployment efficiency.
* Apache 2.0 license
* Multimodal intelligence
* Local hardware deployment
* Massive benchmark leaps
* Efficient MoE architecture
**Models**
* E2B: Mobile efficiency
* E4B: Edge specialist
* 26B MoE: Speed meets intelligence
* 31B Dense: Top-tier performance
* Apache 2.0 license
* Multimodal intelligence
* Local hardware deployment
* Massive benchmark leaps
* Efficient MoE architecture
**Models**
* E2B: Mobile efficiency
* E4B: Edge specialist
* 26B MoE: Speed meets intelligence
* 31B Dense: Top-tier performance
An open-source, theoretical implementation of the Claude Mythos model architecture. The project implements a Recurrent-Depth Transformer (RDT) consisting of three stages: a Prelude, a looped Recurrent Block, and a final Coda. It utilizes switchable attention between Multi-Latent Attention (MLA) and Grouped Query Attention (GQA), alongside a sparse Mixture of Experts (MoE) design to facilitate compute-adaptive reasoning in continuous latent space.
Key technical features include:
* Recurrent-Depth Transformer architecture for implicit chain-of-thought reasoning.
* LTI-stable injection parameters to prevent residual explosion during training.
* Support for multiple model scales ranging from 1B to 1T parameters.
* Integration of Adaptive Computation Time (ACT) or similar halting mechanisms to manage overthinking.
* Use of fine-grained MoE with shared experts to balance breadth and depth.
Key technical features include:
* Recurrent-Depth Transformer architecture for implicit chain-of-thought reasoning.
* LTI-stable injection parameters to prevent residual explosion during training.
* Support for multiple model scales ranging from 1B to 1T parameters.
* Integration of Adaptive Computation Time (ACT) or similar halting mechanisms to manage overthinking.
* Use of fine-grained MoE with shared experts to balance breadth and depth.
Alibaba's Qwen team has open-sourced Qwen3.6-35B-A3B, a sparse mixture-of-experts (MoE) model designed for high performance with low computational costs. While the model possesses 35 billion total parameters, it only activates 3 billion during operation, allowing it to outperform larger dense models in logical reasoning and programming tasks.
Key highlights:
- Uses MoE architecture to achieve high intelligence with minimal activated parameters.
- Demonstrates exceptional multimodal capabilities, particularly in spatial intelligence and visual perception.
- Competes closely with large-scale models like Gemma4-31B and Claude Sonnet 4.5 in specific metrics.
- Integrated into Qwen Studio and available via Alibaba Cloud BaiLian as qwen3.6-flash.
- Supports advanced features like thinking chain retention and seamless integration with AI programming assistants.
Key highlights:
- Uses MoE architecture to achieve high intelligence with minimal activated parameters.
- Demonstrates exceptional multimodal capabilities, particularly in spatial intelligence and visual perception.
- Competes closely with large-scale models like Gemma4-31B and Claude Sonnet 4.5 in specific metrics.
- Integrated into Qwen Studio and available via Alibaba Cloud BaiLian as qwen3.6-flash.
- Supports advanced features like thinking chain retention and seamless integration with AI programming assistants.
The article details the release of Qwen3-Coder-Next, a new 80-billion-parameter open-source large language model (LLM) from Alibaba’s Qwen team. This model is designed for coding tasks and utilizes an ultra-sparse Mixture-of-Experts (MoE) architecture, activating only 3 billion parameters at a time for increased efficiency. It boasts a massive 262,144 token context window and innovative techniques like Gated DeltaNet and Best-Fit Packing to overcome traditional LLM limitations. Qwen3-Coder-Next was trained using an "agentic training" pipeline, learning from real-world coding scenarios and feedback. It supports 370 programming languages and demonstrates competitive performance against leading models like OpenAI’s Codex and Anthropic’s Claude, while also exhibiting strong security features. The release is positioned as a significant advancement in open-weight AI and a challenge to proprietary coding models.
Zhipu AI has released GLM-4.7-Flash, a 30B-A3B MoE model designed for efficient local coding and agent applications. It offers strong coding and reasoning performance with a 128k token context length and supports English and Chinese.
An in-depth look at the architecture of OpenAI's GPT-OSS models, detailing tokenization, embeddings, transformer blocks, Mixture of Experts, attention mechanisms (GQA and RoPE), and quantization techniques.
