---
title: "Welcome Inkling by Thinking Machines"
description: "Thinking Machines Lab has released Inkling and the smaller Inkling-Small family as open multimodal models that accept image, text, and audio inputs. Inkling is described as a roughly 1T-parameter decoder-only Mixture-of-Experts model, with 975B total and 41B active parameters, a 1M-token context window, and training on 45 trillion multimodal tokens. Its architecture combines relative attention, alternating global and sliding-window attention, short convolution, shared experts, hierarchical image patchification, and discretized mel-spectrogram audio processing. The release includes BF16 and NVFP4 checkpoints, speculative MTP layers, day-zero support in Transformers, SGLang, vLLM, and llama.cpp, plus deployment configurations and interactive voice and image demos. Reported benchmarks show Inkling and Inkling-Small achieving strong but varied results across reasoning, coding, vision, audio, factuality, chat, and safety evaluations."
---

# Welcome Inkling by Thinking Machines

[Hugging Face](https://yomu.fyi/company/hugging-face) · ben burtenshaw, merve, Pedro Cuenca, Aritra Roy Gosthipaty, Andres Marafioti · Jul 15, 2026

**Type:** Announcement

## Summary

Thinking Machines Lab has released Inkling and the smaller Inkling-Small family as open multimodal models that accept image, text, and audio inputs. Inkling is described as a roughly 1T-parameter decoder-only Mixture-of-Experts model, with 975B total and 41B active parameters, a 1M-token context window, and training on 45 trillion multimodal tokens. Its architecture combines relative attention, alternating global and sliding-window attention, short convolution, shared experts, hierarchical image patchification, and discretized mel-spectrogram audio processing. The release includes BF16 and NVFP4 checkpoints, speculative MTP layers, day-zero support in Transformers, SGLang, vLLM, and llama.cpp, plus deployment configurations and interactive voice and image demos. Reported benchmarks show Inkling and Inkling-Small achieving strong but varied results across reasoning, coding, vision, audio, factuality, chat, and safety evaluations.

## Context

Thinking Machines Lab released Inkling as a large open model for reasoning across image, audio, and text modalities, with a stated 1M context window and intended use in multimodal applications and domain adaptation through fine-tuning. The post also introduces Inkling-Small for smaller deployment configurations.

## Approach / What changed

Inkling uses a decoder-only multimodal Mixture-of-Experts architecture with 975B total and 41B active parameters. It combines relative attention, a 5:1 sliding-window-to-global-attention pattern, short convolution, routed and shared experts, hierarchical image patchification, and discretized mel-spectrogram audio inputs. The release provides BF16 and NVFP4 variants, speculative MTP layers, support for Transformers, SGLang, vLLM, and llama.cpp, and tested cluster and Inference Endpoints configurations.

## Takeaways

- Inkling uses 256 experts, selecting six routed experts alongside two shared experts that remain active; only 41B of its 975B parameters are active at a given time.
- The BF16 Inkling checkpoint requires 2 TB of VRAM, while the NVFP4 version requires 600 GB. Inkling-Small NVFP4 is listed at 180 GB and can be deployed on one NVIDIA B300 or two NVIDIA H200 GPUs.
- The reported benchmark table shows Inkling scoring 97.1% on AIME 2026 for Inkling-Small and 97.1% for Inkling, while results vary substantially across factuality, agentic, vision, audio, and safety evaluations.

**Tags:** [Architecture](https://yomu.fyi/topic/architecture), [Deployment](https://yomu.fyi/topic/deployment), [LLMs](https://yomu.fyi/topic/llm), [Open Source](https://yomu.fyi/topic/open-source), [Performance](https://yomu.fyi/topic/performance)

- Source: [Hugging Face](https://huggingface.co/blog/thinkingmachines-inkling)
- Source URL: https://huggingface.co/blog/thinkingmachines-inkling
- Ingested by Yomu: 2026-08-29T04:00:57.171Z

[Read original post](https://huggingface.co/blog/thinkingmachines-inkling)