Skip to main content

Vendor

Thinking Machines Lab Models: All 5 Models Compared

Thinking Machines Lab has released 5 models. The strongest is Inkling at rank 71 of 99 language models.

Models55 of them since January 2026
Best rank71Inkling, Rank 71 of 99
Cheapest API$0.30Inkling Small, input per 1M tokens
Most popular61Inkling Small, popularity index out of 100

Summary

Thinking Machines Lab describes itself as an artificial intelligence research and product company. Its stated goal is to make AI systems more widely understood, customizable and generally capable.

In July 2026 it released the open models Inkling and Inkling Small, which process text, images and audio. Inkling 2 Small, the first model of the second generation, followed on 5 November 2026, and Inkling 2 Mini, a looped transformer with 26.4 billion parameters, arrived on 20 November 2026 together with the research checkpoint TML-d24-varloop-research. The weights are under Apache 2.0.

Its own platform Tinker offers the Inkling models for inference and fine-tuning. In March 2026 the company announced a multi-year partnership with Nvidia covering at least 1 gigawatt of Vera Rubin systems, alongside an investment by Nvidia whose amount was not disclosed.

All 5 models from Thinking Machines Lab

The directory also lists announced and retired models. Models with a profile lead to benchmarks, prices and sources.

5 models
  • Released
    Nov 20, 2026
    Status
    Active
    Overall index
    n/a
    API per 1M tokens
    n/a
    Tokens/s
    n/a
    Context
    1M
    Popularity
    n/a
  • Released
    Nov 20, 2026
    Status
    Research preview
    Overall index
    n/a
    API per 1M tokens
    n/a
    Tokens/s
    n/a
    Context
    1M
    Popularity
    n/a
  • Released
    Nov 5, 2026
    Status
    Active
    Overall index
    n/a
    API per 1M tokens
    n/a
    Tokens/s
    n/a
    Context
    n/a
    Popularity
    42
  • Released
    Jul 30, 2026
    Status
    Active
    Overall index
    67.2Rank 76 of 99
    API per 1M tokens
    $0.30 / $1.20
    Tokens/s
    126
    Context
    1M
    Popularity
    61
  • Released
    Jul 15, 2026
    Status
    Active
    Overall index
    71.9Rank 71 of 99
    API per 1M tokens
    $1 / $4.05
    Tokens/s
    107
    Context
    1M
    Popularity
    37

The current Thinking Machines Lab models explained

Inkling 2 Mini

Language model · Nov 20, 2026

Inkling 2 Mini is Thinking Machines Lab's open multimodal looped transformer with 26.4 billion parameters. Its middle 12 of 24 layers run three times. The model processes text and images, supports up to one million context tokens, and is also available through Tinker.

Inkling 2 Small

Language model · Nov 5, 2026

Inkling 2 Small is the first model of Thinking Machines Lab's second generation, released on November 5, 2026. It is a mixture-of-experts model with 594 billion total parameters, 17 billion of them active. It accepts text and images, answers in text, and offers adjustable effort. According to Thinking Machines it scores 31 on the Artificial Analysis Intelligence Index, well above the lab's earlier models.

Inkling Small

Language model · Jul 30, 2026

Inkling Small brings native text, image, and audio handling to a smaller sibling of Inkling. The open model has 276 billion total parameters, activates 12 billion, and places a 1M-token context in a 42-layer sparse MoE with global and local attention.

Inkling

Language model · Jul 15, 2026

Inkling combines native text, image, and audio handling with 975 billion total parameters and 41 billion active per token. Thinking Machines Lab places its 1M-token context in a 66-layer sparse MoE with global and local attention and names Inkling Small as another model in the family.

Which Thinking Machines Lab model fits?

Derived from the overall index, the API price and the measured speed of the current models.

The strongest choice

Inkling

The vendor's highest overall index

Overall index
71.9 (Rank 71 of 99)
API per 1M tokens
$1 / $4.05
Speed
107 tokens/s

Value for money

Inkling Small

Costs 30% of Inkling, overall index 67.2 instead of 71.9

Overall index
67.2 (Rank 76 of 99)
API per 1M tokens
$0.30 / $1.20
Speed
126 tokens/s

Thinking Machines Lab in the overall index

The 20 best language models and every Thinking Machines Lab model in the index, with the Thinking Machines Lab models highlighted. “All” and the filter show every model in the index.

Overall index (Gradually Meta-Score v6)

22 of 99 models

Gemini 4 Neonxhigh127.1
GPT-6 Astramax119.6
GPT-6.2 Solmax119.2
GPT-6.1 Solmax117.3
Muse Wondermax113.5
GPT-5.6 Solmax109.5
GPT-6 Solmax108.3
Inklingxhigh71.9
Inkling Small67.2

Where does Thinking Machines Lab stand?

Models among the best 20 in the overall index

Anthropic6 models, 4 of them among the best ten
OpenAI6 models, 3 of them among the best ten
Google4 models, 3 of them among the best ten
Meta3 models, 0 of them among the best ten
Moonshot AI1 models, 0 of them among the best ten
Thinking Machines Lab0 models, 0 of them among the best ten

Best Thinking Machines Lab model per area

Finance
Rank 10 of 115
Inkling 2 Small
Legal
Rank 12 of 83
Inkling 2 Small
Health
Rank 15 of 72
Inkling 2 Small
Knowledge
Rank 28 of 31
Inkling
General
Rank 29 of 72
Inkling 2 Small
Coding
Rank 40 of 126
Inkling 2 Small
Reasoning
Rank 61 of 87
Inkling
Agents
Rank 73 of 111
Inkling Small

Vendors compared

  • Models
    35
    Language models
    35
    Image models
    0
    Video models
    0
    Audio models
    0
    Decision models
    0
    Specialist models
    0
    Top 10
    4
    Open weights
    0
    New in 2026
    15
    Best language model
    Claude Fable 5.2 (Rank 1)
  • Models
    69
    Language models
    41
    Image models
    5
    Video models
    7
    Audio models
    11
    Decision models
    1
    Specialist models
    4
    Top 10
    3
    Open weights
    15
    New in 2026
    32
    Best language model
    Gemini 4 Carbon (Rank 2)
  • Models
    74
    Language models
    50
    Image models
    8
    Video models
    2
    Audio models
    11
    Decision models
    1
    Specialist models
    2
    Top 10
    3
    Open weights
    4
    New in 2026
    36
    Best language model
    GPT-6.2 Astra (Rank 4)
  • Models
    35
    Language models
    30
    Image models
    1
    Video models
    1
    Audio models
    1
    Decision models
    0
    Specialist models
    2
    Top 10
    0
    Open weights
    27
    New in 2026
    9
    Best language model
    Muse Wonder Full (Rank 11)
  • Models
    12
    Language models
    12
    Image models
    0
    Video models
    0
    Audio models
    0
    Decision models
    0
    Specialist models
    0
    Top 10
    0
    Open weights
    11
    New in 2026
    7
    Best language model
    Kimi K3.1 Pro (Rank 16)
  • Thinking Machines Lab
    Models
    5
    Language models
    5
    Image models
    0
    Video models
    0
    Audio models
    0
    Decision models
    0
    Specialist models
    0
    Top 10
    0
    Open weights
    5
    New in 2026
    5
    Best language model
    Inkling (Rank 71)

Gaps

No image models, video models or audio models in the directory
1 model only with limited access or as a preview

Released Thinking Machines Lab models over time

By release date. Grey: retired models.

Releases in the last twelve months

0Dec 25
0Jan 26
0Feb 26
0Mar 26
0Apr 26
0May 26
0Jun 26
2Jul 26
0Aug 26
0Sep 26
0Oct 26
3Nov 26

5 releases since January 2026, by release date.

API prices

Prices per 1M tokens for the current language models, according to Thinking Machines Lab or the provider linked in each model profile.

Plans from other providers with Thinking Machines Lab models

Plans from other providers that include at least one current model from Thinking Machines Lab. Monthly list prices in the provider's currency; taxes may apply. Price sorting uses German prices including VAT.

5 plans
  • Command Code
    Plan
    Gowith credits
    Per month
    $1
    Included models
    Inkling Small
  • Command Code
    Plan
    GOATwith credits
    Per month
    $10
    Included models
    Inkling Small
  • Command Code
    Plan
    Prowith credits
    Per month
    $20
    Included models
    Inkling Small
  • Command Code
    Plan
    Max 10×with credits
    Per month
    $100
    Included models
    Inkling Small
  • Command Code
    Plan
    Max 20×with credits
    Per month
    $200
    Included models
    Inkling Small

Frequently asked questions about Thinking Machines Lab

What Thinking Machines Lab models are there?

The directory lists 5 Thinking Machines Lab models: 5 language models. It counts every model Thinking Machines Lab ever released, including retired ones. 4 models are classified as active, including Inkling 2 Mini, Inkling 2 Small and Inkling Small.

Which Thinking Machines Lab model is the best?

Inkling. It is at rank 71 of 99 in the overall index, which combines independent benchmarks from several sources.

What is the latest Thinking Machines Lab model?

Inkling 2 Mini, released on Nov 20, 2026. The timeline below shows every release by date.

How much do Thinking Machines Lab models cost?

Through the API per 1M tokens (input / output): Inkling Small $0.30 / $1.20 and Inkling $1 / $4.05. The prices of all 2 current language models are in the price table.

Is Thinking Machines Lab better than ChatGPT?

No, not in the overall index. The best OpenAI model, GPT-6.2 Astra (rank 4 of 99), is ahead of Inkling (rank 71 of 99) (71.9 against 126.6 points), and the lead holds up in our calculation. Single tasks can differ.

Can I run Thinking Machines Lab models locally?

Yes, 5 of the 5 models have open weights. Each model profile lists how much memory it needs.

Which Thinking Machines Lab model is the fastest?

Inkling Small with 126 tokens per second according to OpenRouter. A typical model manages 72.

How large is the context window of the Thinking Machines Lab models?

Up to 1 M tokens (≈ 1,500 A4 pages), which 4 of the current models reach, including Inkling 2 Mini. Every model's value is in the model list.