What hardware do you need to run Inkling Small?
Inkling Small at UD-Q4_K_M is 163 GB of weights, with a context ceiling of 1024k tokens. Not yet rated: released after our last ratings pass. Thinking Machines' open weights. 'Small' is relative to a 975B sibling — this still needs 256 GB.
How good is it, really?
On the Artificial Analysis Intelligence Index v4.3 it scores 26, which puts it in the Haiku-class band. In the same band as Anthropic's cheap, fast tier. Every current OpenAI model scores above this band. Score source. See the whole table.
- Summarising — not rated
- Translation — not rated
- Everyday coding — not rated
- Reasoning & maths — not rated
- Agentic work — not rated
What it costs either way
Renting the same model costs $0.45 per million input tokens and $1.2 per million output (OpenRouter, cheapest active endpoint, checked 2026-09-09). Buying a machine only beats that if you use it hard enough, for long enough, that the hardware price divides down below the rental bill.
Machines that run it
| Machine | Price | Speed at 32k | Pay-back | |
|---|---|---|---|---|
| Mac Studio M5 Ultra, 256GB | $10,799 | 43 tok/s estimated | Pays back in 124 years | Run the numbers |
One machine per family, cheapest first. Speeds are measured where a public benchmark exists and estimated from memory bandwidth otherwise; the calculator says which for any configuration.
The specifics
- Parameters
- 266B, of which 12B are active per token
- Quantisation
- UD-Q4_K_M
- Weights on disk
- 163 GB
- KV cache
- 1.0 GB at 32k context — 35 of 42 layers use a 512-token window. Published only as an unsloth dynamic quant, which allocates bits differently from a plain Q4_K_M.
- Maximum context
- 1024k tokens (1M (model_max_length; the config states no max_position_embeddings))
- Licence
- Apache 2.0
- Sources
- source 1, source 2