Strata 0.2.19 Adds Inkling 2 Mini Reasoning Model That Runs on a 4 GB Graphics Card

Published by

Strata 0.2.19 has been released, introducing the Inkling 2 Mini reasoning model, which can operate on a graphics card with as little as 4 GB of memory. This model, developed by Thinking Machines Lab, features a unique structure that allows it to re-process its middle layers multiple times, effectively enhancing its reasoning capabilities. The release also includes support for NVIDIA Blackwell FP4 and improvements in the project's kernel, all without needing environment variable configurations. Strata's rapid development cycle and transparency in performance labeling continue to establish its reputation in the field of AI modeling



Strata 0.2.19 Adds Inkling 2 Mini Reasoning Model That Runs on a 4 GB Graphics Card

Strata 0.2.19 just launched with native support for Thinking Machines Lab's Inkling 2 Mini, a 26.4-billion-parameter "looped transformer" that reasons longer internally by re-running its middle layers. The model fits on hardware as small as a 4 GB graphics card through three labeled quantization packs ranging from 10.47 GB to 19.35 GB, and Strata claims it reaches up to 250 tokens per second on an RTX 5090. The release also adds NVIDIA Blackwell FP4 support and broad Qwen speed boosts while continuing the project's from-scratch kernel rewrite, all with zero environment variables required. It arrives hours after 0.2.18 made Tencent's 220-billion-parameter Hy4-Flash run on a single PC, reinforcing Strata's reputation for fast iteration and unflinchingly honest quality labeling.

Strata 0.2.19 Adds Inkling 2 Mini Reasoning Model That Runs on a 4 GB Graphics Card @ Linux Compatible