作者： Jakub Rusinowski · 最后更新： 2026年7月21日
PREVIEW (July 2026). Thinking Machines Lab's first open-weights model — and the first ~1-trillion-parameter open model to natively accept text, image and audio input with a 1M-token context window. Inkling is a sparse multimodal Mixture-of-Experts transformer: 975B total parameters but only ~41B active per token (256 experts, top-6 routed plus 2 always-on shared experts). It was trained on 45 trillion tokens of text, images, audio and video, and reasons natively across modalities with seven selectable effort levels (none → max). Ships in full BF16 (~2 TB VRAM, Hopper+) and a well-calibrated NVFP4 4-bit build (~600 GB, Blackwell), both with MTP speculative-decoding layers for faster inference. Released under Apache 2.0 and pitched as a customization base to fine-tune via the company's Tinker platform — "built for customization, not leaderboard dominance." Datacenter-class hardware required; most users will reach it through a hosted API. Specs from launch coverage and the Hugging Face model card — verify before relying on them.
| Inkling (NVFP4) | Min 600 GB VRAM · NVFP4 (4-bit) · 1,000,000 ctx · |
| Inkling (BF16) | Min 2000 GB VRAM · BF16 (full precision) · 1,000,000 ctx · |
Install Ollama then run: ollama run
Minimum VRAM: 600 GB. For best results use Q4_K_M quantization.
Inkling needs about 600 GB VRAM at NVFP4 (4-bit) quantization for its smallest variant. Variants: Inkling (NVFP4) (600 GB, NVFP4 (4-bit)); Inkling (BF16) (2000 GB, BF16 (full precision)). On Apple Silicon, unified memory counts toward this requirement.
Inkling's smallest variant needs about 600 GB, which exceeds a single RTX 4090 (24 GB). Use multiple GPUs, a higher-VRAM card, or Apple Silicon with large unified memory.
Q4_K_M is the best balance of quality and VRAM for Inkling in most cases. Choose Q8_0 for near-lossless quality if you have spare VRAM, or smaller quants (Q3/Q2) only when memory is tight.
Install Ollama, then run: ollama run . This downloads Inkling and starts a local, OpenAI-compatible endpoint — no internet connection is needed after the initial download.