Thinking Machines Launches 276B Inkling-Small, Rivaling Larger Models with 12B Active Params
arena · x · 2026-08-01
Thinking Machines Lab has released Inkling-Small, a new open-weights Mixture-of-Experts (MoE) model. It features 276B total parameters but only 12B active parameters, achieving performance comparable to much larger models at a fraction of the compute cost.
In Chatbot Arena's preliminary AutoEval, Inkling-Small scored 1431 points, ranking #88 overall and #21 among open models. It stands neck-and-neck with compute-heavy open models like Minimax M2.7, Nemotron 3 Ultra, and Deepseek V4 Flash. The model supports native reasoning over audio and images with a context window of up to 1M tokens.
More from Models
- Experiment: Prompting Claude to Code a Procedural Bone and Skin Animation System — chongdashu · 2026-08-01
- Comparing 18 Major LLM APIs: Costs Vary by Over 100x for the Same Workload — mentorperplexed · 2026-08-01
- User Reports Claude Opus 5 Feels Janky and Delivers a Worse Experience — vivekhaldar · 2026-08-01
- Fable's AI Safety Filter Constantly Triggers on Benign Content — dreamwieber · 2026-08-01
- V4-Flash Prioritizes Agents Over Peak STEM Reasoning Performance — teortaxesTex · 2026-08-01
- DeepSeek-V4-Flash Inference Blocked: vLLM Lacks Support for New confidence_head — teortaxesTex · 2026-08-01