Inkling
Thinking Machines · released Jul 14, 2026 · thinkingmachines/Inkling
- Type
- Open weightsApache License 2.0
- Params
- 952B
- Context
- 1M
about 786K words of context
Our take
Written Aug 2, 2026Inkling is a 952.4-billion-parameter downloadable model from Thinking Machines with a one-million-token request limit and broad multimodal input. Released in mid-2026 under a permissive Apache licence, it excels at coding and mathematics while its creative writing and agentic performance lag behind.
Choose this when you need a very large downloadable model under a permissive licence, or for long-document work at one million tokens. It suits coding tasks best — that is where its benchmark scores peak — and handles text, image and audio input with text output. Skip it if you need strong creative writing, reliable agentic behaviour, or consistent speed across providers.
The case for it
- 952.4 billion total parameters, the largest we have recorded under Apache License 2.0.
- Coding is its standout skill: 53.4 points above its general chat score on the same benchmark family.
- Strong mathematical reasoning at 1478.8964 on the Arena Maths benchmark.
- One-million-token request limit supports long documents and extended conversations.
The case against it
- Creative writing trails its other capabilities by 54.9 points below general chat and 108.3 below coding on the same benchmark family.
- Agentic performance is negative on the one measured task, scoring -0.0571.
- Throughput varies sharply by provider: the fastest measured is 2.4 times the slowest.
How good is it?
IntelligencePuzzles, maths, exam questions
Arena Text (overall)45th of 143 · 1440.8
CodingWriting and fixing code on its own
Arena Coding39th of 143 · 1497.1
AgenticPlanning, calling tools, staying on task
Arena Agent (IPS)30th of 36 · −0.057
WritingWe do not rate this
Two boards come close and neither tests writing: Arena Creative Writing asks people which of two replies they prefer, and LiveBench Language tests whether a model understood a passage. So we show where Inkling placed and give it no mark out of five.
These tests check whether a model follows instructions — a precondition for all the work above, but not a measure of how well that work is done, which is why they get no rating.
Every published score for this model16 scoresEvery figure we hold, from 16 boards, with who ran it and a link to the source — including the boards no rating above is built on.
Can you run it yourself?
- Fits in memory
- weights load entirely on the card
- Spills to system RAM
- some weights offload; much slower
- Too large
- will not load even with offload
- est
- size is calculated; the verdict could change by 10%
GeForce RTX 4090 · 24 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Apple M1 Pro (16-core GPU) · 32 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Radeon RX 7900 XT · 20 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Memory use by level
Against a 24 GB card.
All 3 sizes here are calculated, not measured. We hold no measured file for this model, so each size comes from the parameter count and every verdict above inherits that uncertainty.
Check against your own machine → · All 71 devices, with every size →
Or rent it from someone else
Cheapest of 5 live listings. Picked at the widest standard context we hold, within one quantisation slice, so the numbers beside it are a price one host actually charges.
- per 1M tokens
- $1.00 in / $4.05 out
- Context served
- 1M
- Throughput
- Not measured
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| OpenRouter | $1.00 / $4.05 | 1M | not measured | Unknown | Unknown | Unknown |
| DeepInfrafp8 | $0.95 / $4.05 | 524K | not measured | Unknown | Unknown | Unknown |
| DeepInfrafp8 | $0.95 / $4.05 | 524K | 146 tok/s | No | No | Confirmed |
| Together AI | $1.00 / $4.05 | 524K | 92 tok/s | No | No | Confirmed |
| Basetenfp8 | $1.00 / $4.05 | 1M | 91 tok/s | No | No | Confirmed |
DeepInfra
Together AIAcross the 5 listings we hold: 3 say they do not train on prompts, 0 say they do and 2 do not say. 3 appear in the zero-retention registry we check; the rest are unknown to us rather than confirmed either way.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
- ✓
- Supported
- ✗
- Not supported
- Not published
- host gave no parameter list
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| OpenRouter | ✓ | ✓ | ✗ |
| DeepInfrafp8 | |||
| DeepInfrafp8 | ✗ | ✓ | ✗ |
| Together AI | ✓ | ✗ | ✗ |
| Basetenfp8 | ✓ | ✗ | ✗ |
Tool calling: 3 of 5 listings say yes, 1 says no, 1 publishes no parameter list. JSON output: 2 of 5 listings say yes, 2 say no, 1 publishes no parameter list. Strict schema: 0 of 5 listings say yes, 4 say no, 1 publishes no parameter list.
When we formed this view
Dates behind this page
Prices last checked 14h ago
What we do not know about this model yet
- We hold no measured file for it, so all 3 sizes on this page are calculated from the parameter count.
- 1 of 5 listings publish no parameter list, so what their API accepts is unknown to us.
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
- 2 of 5 listings do not say whether they train on prompts.
Licence and identifiers
What the licence allowsApache License 2.0, what it allows commercially, and the identifiers you need to pull this model — its Hugging Face repo, our slug and a machine-readable card.
Licence
Apache License 2.0
Fully permissive: commercial use, redistribution, and derivatives allowed. Requires attribution and a copy of the license. Includes an express patent grant.
Identifiers
- Hugging Face
- thinkingmachines/Inkling
- Architecture
- Dense
- Modality record
- text+image+audio->text
- Catalogue slug
- thinkingmachines-inkling