Size
594B
17B active — only part of the model works on each answer, but all of it has to fit in memory
Inkling 2 Small
Context
1M
the longest conversation the model is built to hold
Lab
Thinking Machines
the lab that released it
Released
Nov 12, 2026
Licence: Apache 2.0 — free for any use, including commercial. Just keep the licence notice.
Worth knowing
Thinking Machines' second-generation model: a 48-layer MoE routing each token to 8 of 384 experts, with a 292M-parameter vision encoder, image input and variable reasoning effort. Launched Nov 5, 2026 on Tinker with closed weights; the full weights landed on Hugging Face on Nov 12, 2026 under Apache 2.0, as promised. Two checkpoints: the native NVFP4 build, whose card requires a GPU cluster with at least 500GB of aggregated VRAM (the card's examples: 2x B300 or 4x H200), and an MXFP8 build at at least 800GB (4x B300 or 8x H200); the card recommends NVFP4. Serving via SGLang, vLLM, TokenSpeed or Unsloth. Artificial Analysis Intelligence Index 31 at launch — the lab's figure.
What it needs
Full precision (FP16)
1,188 GB
the weights, uncompressed
8-bit
594 GB
the weights, compressed to 8-bit
4-bit
297 GB
the weights, compressed to 4-bit
How we work this out: memory = the model's total parameters × bytes per weight — 2 bytes at full precision (FP16), 1 byte at 8-bit, half a byte at 4-bit. Weights only — add room for the conversation memory (KV cache) and the serving software. A GPU counts as running a build when its memory covers the weights plus about 20% headroom. Real figures move with context length and software.
GPUs that can run it
Ranked by how well the model runs, then by price.
No single GPU can run Inkling 2 Small — even its 4-bit build needs 297GB, plus room to work, and no single card holds that. Needs at least 2× B300 by memory at 4-bit — a memory fact, not a performance promise. The prices board shows the biggest cards and what each is short by, with the single-GPU and node prices a setup like this is built from.
The minimum setup, by memory
Needs at least 2× B300 by memory at 4-bit — a memory fact, not a performance promise.
No single card runs Inkling 2 Small alone at this size — it is a multi-GPU job. The board lists the cards, nodes and racks a setup is built from, with node offers shown by their total price.
See the GPUs for Inkling 2 Small →
Checked Nov 15, 2026
Select a GPU See these GPUs on the prices board, ranked for this model.