Inkling — local AI model by Thinking Machines
Written by Jakub Rusinowski · Last updated
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.
Variants
The smallest Inkling variant needs about 589 GB of VRAM at NVFP4 (4-bit) — quantized weights plus framework overhead, before any KV cache.
| Model | VRAM at Q4 | VRAM | Context | Run it |
|---|---|---|---|---|
| Inkling (NVFP4) → 1T | ~589.5 GB | 1,000,000 | cloud-hosted tag | |
| Inkling (BF16) → 1T | ~589.5 GB | 1,000,000 | cloud-hosted tag |
Memory is quantized weights plus overhead at Q4_K_M, from the same engine as the GPU & VRAM checker.
How to run Inkling locally
Install Ollama, then pull the tag.
No Ollama library entry. Thinking Machines' documented serving routes are SGLang and vLLM.
Pick a size above for its own VRAM figure, speed estimate and install command.
Licence
Commercial use permitted. No usage restrictions beyond attribution.
Applies to: Inkling (NVFP4), Inkling (BF16)Can I run Inkling on my GPU?
Inkling — frequently asked questions
How much VRAM does Inkling need?
Inkling needs about 589 GB VRAM at NVFP4 (4-bit) quantization for its smallest variant. Variants: Inkling (NVFP4) (589 GB, NVFP4 (4-bit)); Inkling (BF16) (589 GB, BF16 (full precision)). On Apple Silicon, unified memory counts toward this requirement.
Can I run Inkling on an RTX 4090 (24 GB)?
Inkling's smallest variant needs about 589 GB, which exceeds a single RTX 4090 (24 GB). Use multiple GPUs, a higher-VRAM card, or Apple Silicon with large unified memory.
What quantization should I use for Inkling?
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.
How do I run Inkling with Ollama?
Inkling has no local Ollama tag — the published tag is cloud-hosted, so running it sends your prompts to a hosted GPU rather than your own machine.