This is a GGUF-quantized version of the **Krea-2-Turbo** diffusion transformer, produced with **ByteShape's ShapeLearn**, which learns the optimal datatype per tensor to maintain high quality even at very low bitlengths. Five sizes are available, from **14.31 GB (8.93 bpw)** down to **6.26 GB (3.91 bpw)**. All variants generate high-quality images. These files run in **ComfyUI** via the ComfyUI-GGUF extension.
This is a GGUF-quantized version of the **Qwen-Image-2512** diffusion transformer, produced with **ByteShape's ShapeLearn**, which learns the optimal datatype per tensor to maintain high quality even at very low bitlengths. Six sizes are available, from **16.87 GB (6.61 bpw)** down to **7.77 GB (3.04 bpw)**. All variants generate high-quality images. > Looking for more speed? We also publish a **vLLM-Omni** build of the same model with optimized Humming kernels, roughly **2-3x faster** per step. See byteshape/Qwen-Image-2512-Humming.
This is a GGUF-quantized version of Qwen3.6-35B-A3B produced with **ByteShape's ShapeLearn**, which learns the optimal datatype per tensor to maintain high quality even at very low bitlengths. To learn more about ShapeLearn and to see detailed benchmarks across GPUs, CPUs, and even the Raspberry Pi, please visit our blog. Pick a model from the tables below and click **Get llama.cpp command** to get a ready-to-run command with all the correct sampling parameters for this model.
This is a GGUF-quantized version of Qwen3.6-35B-A3B with **Multi-Token Prediction (MTP)** support, produced with **ByteShape's ShapeLearn**, which learns the optimal datatype per tensor to maintain high quality even at very low bitlengths. MTP allows the model to predict multiple tokens per forward pass, accelerating decoding on GPUs by accepting speculative tokens that match what the base model would have produced. These models are optimized for **GPU-only** inference. Pick a model from the table below and click **Get llama.cpp command** to get a ready-to-run command with all the correct sampling parameters for this model.
This is a GGUF-quantized version of Qwen3.8-27B produced with **ByteShape's ShapeLearn**, which learns the optimal datatype per tensor to maintain high quality even at very low bitlengths. This release replaces our earlier ShapeLearn-Lite quantizations of Qwen3.8-27B. Every model here has been through our **full multi-benchmark evaluation suite**, covering math, coding, general knowledge, instruction following and agentic tool use in both thinking and instruct modes, scored against the BF16 baseline. To learn more about ShapeLearn and to see detailed benchmarks across GPUs, please visit our blog.
No-Thinking · PI Tune · MTP · GGUF 🪐 Qwen3.6-27B-MTP-pi-tune A 4-bit QLoRA SFT Multi-Token Prediction tune of Qwen3.6-27B fine-tuned for no-thinking agentic coding through a PI-style harness. Packaged as llama.cpp-compatible GGUF for local agent loops. 🧠 27B Dense Foundation 🚫 No-Thinking Tuned ⚡ MTP Speculative Decoding 🛠️ Coding · DevOps · Agents 📦 llama.cpp GGUF 🖼️ Multi-Modal Compatible 🪟 256k Native · 1M Max Context > [!TIP] > For the strongest Pi-style coding-agent behavior, use the reasoning-trained release: `bytkim/Qwen3.6-27B-MTP-pi-reasoning-GGUF`. See the technical writeup for the broader evaluation context. This no-thinking tune remains useful when you specifically want the lower-latency direct / instruct path.
**Qwen3.8-27B-pi** builds on Qwen3.8-27B for coding work in the Pi agent harness. It is tailored to the loop of reading a repository, editing files, running tools, and responding to their feedback. The aim is more completed work with less generated text, while retaining the base model's familiar interface. Fine-tuned for the Pi agent harness, the model is designed to work through coding tasks as an iterative process: inspect the code, make a change, check the result, and use tool feedback to guide the next step. Adjustable reasoning effort lets you balance responsiveness with deeper problem-solving, from focused edits to more involved debugging and implementation work. - **Built for Pi:** Fine-tuned for the everyday coding loop—reading repositories, editing files, running tools, and working through feedback in the Pi harness—with an emphasis on turning plans into working, checked implementations. - **Curated coding experience:** Supervised fine-tuning on filtered, successful Pi sessions helps the model learn complete coding workflows, rather than isolated answers or code snippets, including how to adapt to existing environments and check results against task requirements. - **Task quality matters:** Development also included reviewing, repairing, and exploring more demanding coding tasks—with a focus on clear requirements and checks that distinguish working solutions from incorrect ones. - **Refined through reinforcement learning:** A second training stage pairs verified task outcomes with a custom reasoning-efficiency reward built on GRPO, encouraging successful low- and medium-effort solutions to reason more economically while leaving xhigh focused on correctness. - **Available in practical formats:** Choose a deployment format that fits your hardware, including GGUF quantizations calibrated using complete Pi coding sessions.
Adjust provider, category or input/output filters. New models are added continuously.
COST EXPLORER
Great models. Clear costs.
Cost analysis
Adjust monthly token usage to compare estimated API costs.
1 M = 1 million tokens · USD Compares base text-token rates within the same channel
Uses collected, comparable USD-per-million-token rates. Snapshots exclude cache discounts, tiers, extra reasoning output and taxes. Check current source prices.
MODEL HANDBOOK
Understand before you choose
Selection guide & FAQ
How should API prices be compared?+
Separate input, output and cache rates, then estimate the same workload. Image, audio and video rates may use items, duration or resolution rather than tokens.
Do open weights mean free commercial use?+
Not necessarily. Check each version’s license, commercial restrictions and policies. Self-hosting also incurs compute and maintenance costs.
Is the top-ranked model the best for me?+
Rankings depend on tasks, versions and methods. Compare accuracy, latency, reliability and costs on your own workload.
