Developer docs
API playgroundTry for free, no card

Search company profiles

Thinking Machines Lab

Full company profile

uuid0000087

Namestring
Thinking Machines Lab
Legal namestring
Thinking Machines Lab, Inc.
Company typeenum
Private
Founded yearint
2025
Descriptiontext

Thinking Machines Lab is a private, AI research and product company founded in February 2025 in San Francisco by former OpenAI CTO Mira Murati alongside a group of OpenAI alumni co-founders. The company's stated mission is to make AI systems more widely understood, customizable, and generally capable, with an explicit emphasis on human-AI collaboration rather than fully autonomous agents. Its product surface centers on two pillars: Tinker, a LoRA-based training and fine-tuning API launched in October 2025 that abstracts distributed compute for researchers running RL and supervised learning experiments across a broad lineup of open-source models (Qwen, GPT-OSS, DeepSeek, Moonshot Kimi, NVIDIA Nemotron); and TML-Interaction-Small, a proprietary 276B-parameter mixture-of-experts multimodal model (12B active) released as a research preview in May 2026 that natively processes audio, video, and text in 200ms time-aligned micro-turns and achieves state-of-the-art FD-bench v1.5 and turn-taking latency scores.

The underlying technology stack combines proprietary model architecture (encoder-free early fusion, dMel audio embedding, batch-invariant inference kernels) with significant published research contributions on the Connectionism blog covering LoRA fine-tuning methodology, on-policy distillation (demonstrating 9–30x compute cost reductions), and deterministic inference. The company has locked in frontier compute capacity through a multi-year, gigawatt-scale NVIDIA Vera Rubin deployment (targeting early 2027) and a multi-billion-dollar Google Cloud agreement on GB300-powered A4X Max VMs, with NVIDIA also making a significant equity investment.

Thinking Machines Lab's commercial model is primarily API-first and usage-based, with Tinker priced per million tokens across Prefill, Sample, and Train operations plus $0.10/GB-month storage. Early customers skew academic and research-oriented (UC Berkeley, Princeton, Stanford, Redwood Research), with enterprise revenue anchored by a $471.7M non-cancelable cloud-infrastructure contract with Boost Run and adjacent infrastructure services supplied to customers like RadixArk. Distribution runs through self-serve developer signup supplemented by direct enterprise agreements and cloud marketplace channels; revenue remains undisclosed and the company has been capital-intensive rather than profit-oriented since founding, with a $2B seed round at a $12B valuation in July 2025 underwriting ongoing buildout.

Short descriptiontext

Thinking Machines Lab is a San Francisco-based AI research and product company building frontier multimodal interaction models (TML-Interaction-Small) and a fine-tuning API (Tinker), serving researchers, developers, and enterprises with customizable, real-time human-AI collaboration systems.

Operating statusenum
Operating
Ownership categoryenum
Headcount rangeband
101–250
akta.pro rankint
HeadquartersSan Francisco, United States
HQ citystring
San Francisco
HQ countrystring
United States
HQ regionstring
North America
Markets served

Serves global market

Offices2 records

Each record includes

City, Country, Type, Description, Source

Keyword5 values
multimodal AI models, AI fine-tuning API, foundation model research, AI infrastructure services, open-source AI training
Industry4 codes
PRIMARYHDAAACAC
Fine-Tuning, Adaptation & Custom Model Training (PEFT/LoRA/RLHF)
Hardware, Technology & Telecom/Artificial Intelligence & Machine Learning/Foundation Models & Generative AI (LLMs, multimodal)
SECONDARYHDAEANAB
Model Development & Training Platforms (AutoML, Notebooks, Feature Stores)
Hardware, Technology & Telecom/Enterprise Platforms & Databases/Artificial Intelligence Platforms for Enterprise
SECONDARYHDAEANAD
LLMOps & Generative AI Platforms (Prompt/Agent Orchestration, RAG)
Hardware, Technology & Telecom/Enterprise Platforms & Databases/Artificial Intelligence Platforms for Enterprise
SECONDARYHDAAAAAI
AI Compiler, Runtime & Kernel Optimization Software (CUDA/ROCm/XLA, graph compilers)
Hardware, Technology & Telecom/Artificial Intelligence & Machine Learning/AI Infrastructure & Compute (GPUs/ASICs, accelerators)
NAICS code3 codes
  • Scientific Research and Development Services5417
  • Computer Systems Design and Related Services54151
  • Other Computer Related Services541519
SIC code1 code
  • Services-Computer Programming Services7371
Product category
Foundation Models / AI Infrastructure
GTM motion1 record

Each record includes

Type, Description, Source

Revenue model3 records
1Tinker API Consumption
TypeUsage Based
Description

Usage-based revenue from token consumption (Prefill, Sample, Train pricing per million tokens across multiple model families including Qwen, GPT-OSS, DeepSeek, Kimi, NVIDIA Nemotron). Storage charged at $0.10/GB-month. Pre-paid credits system with promotional grants available.

thinkingmachines.ai
2Enterprise Service Agreements
TypeSubscription Recurring
Description

Pre-paid credits for Services (paid API/hosted model access); Enterprise Agreements possible for enterprise customers with proprietary terms; non-refundable fees with 1-year expiration on pre-paid credits

thinkingmachines.ai
3Multi-Billion-Dollar Cloud/Compute Contracts
TypeManaged Services
Description

Multi-year infrastructure deals with Google Cloud (single-digit billions) and NVIDIA (gigawatt-scale Vera Rubin deployment) for AI model training and deployment capacity

siliconangle.com
Marketing channels8 records

Each record includes

Title, Type, Stage, Description, Source

Distribution channels3 records

Each record includes

Title, Type, Scope, Target buyer, Description, Source

Cost components5 values
Infrastructure, Personnel, Technology or R&D, Marketing or Sales, Operations
Pricing details7 tiers
1NVIDIA Nemotron-3-Nano-30B-A3B-BF16 - small MoE tier at $0.13 Prefill / $0.33 Sample / $0.40 Train per million tokens (50% limited-time discount)
ModelUsage-basedBilling cadencePay-as-you-go
Notes

64K context length; limited-time 50% discount applied

thinkingmachines.ai
2NVIDIA Nemotron-3-Super-120B-A12B-BF16 - mid-tier model at $0.38 Prefill / $0.96 Sample / $1.16 Train per million tokens (50% discount)
ModelUsage-basedBilling cadencePay-as-you-go
Notes

64K and 256K context options; 50% limited-time discount

thinkingmachines.ai
3NVIDIA Nemotron-3-Ultra-550B-A55B-BF16 - largest model at $1.66-$3.32 Prefill / $4.15-$8.30 Sample / $4.98-$9.96 Train per million tokens (50% discount)
ModelUsage-basedBilling cadencePay-as-you-go
Notes

64K and 256K context options; 50% limited-time discount

thinkingmachines.ai
4GPT-OSS-20B at $0.12 Prefill / $0.30 Sample / $0.36 Train per million tokens (32K context)
ModelUsage-basedBilling cadencePay-as-you-go
Notes

Compact open-source model pricing

thinkingmachines.ai
5Qwen3.5-397B-A17B - $2.00-$4.00 Prefill / $5.00-$10.00 Sample / $6.00-$12.00 Train per million tokens
ModelUsage-basedBilling cadencePay-as-you-go
Notes

Large MoE model; 64K and 256K context options

thinkingmachines.ai
6Kimi K2.5 - $1.47 Prefill / $3.66 Sample / $4.40 Train per million tokens; retiring July 12
ModelUsage-basedBilling cadencePay-as-you-go
Notes

Retiring July 12 per deprecation notice; 32K and 128K context options

thinkingmachines.ai
7Storage: $0.10 per GB-month across all tiers
ModelUsage-basedBilling cadencePay-as-you-go
Notes

Storage charged separately from token usage

thinkingmachines.ai
GTM typeB2B
B2B
Offering typeSoftware
Software
Brand1 of 3 records shown
1Tinker
Description

A training API for researchers providing fine-tuning of open-source models with LoRA; handles scheduling, tuning, resource management, and infrastructure reliability. Includes a Tinker cookbook and a public X account @tinkerapi.

thinkingmachines.ai
+2 more records
Core offering1 text field

Thinking Machines Lab builds and sells AI research products, primarily the Tinker fine-tuning API for researchers and developers (enabling LoRA-based training on open-source models like Qwen, GPT-OSS, DeepSeek, and NVIDIA Nemotron) and proprietary multimodal interaction models (TML-Interaction-Small) that natively handle real-time audio, video, and text collaboration with 200ms micro-turn latency. The company sells usage-based API access to its training platform and grants research preview access to its frontier interaction models.

Differentiator
Functional benefit
Problem solved
Product overview1 text field

Thinking Machines Lab offers a platform-plus-modules product portfolio anchored on two main pillars: (1) Tinker, a managed fine-tuning API launched in October 2025 that lets researchers and developers apply LoRA-based supervised learning and RL training to a broad lineup of open-source models (Qwen, GPT-OSS, DeepSeek, Moonshot Kimi, NVIDIA Nemotron) without managing GPU infrastructure; and (2) TML-Interaction-Small, a proprietary 276B-parameter multimodal interaction model introduced as a research preview in May 2026 that natively handles full-duplex audio, video, and text interaction in 200ms micro-turns. The Connectionism research blog and News page round out the public-facing surface, publishing technical posts (LoRA Without Regret, On-Policy Distillation, Defeating Nondeterminism in LLM Inference), product announcements, and research grant programs (Interactivity Research Grants, Tinker Research and Teaching Grants) that bridge research outputs to the Tinker training platform and the company's frontier interaction models.

Product and service3 records
1Tinker
CategoryAI Fine-tuning API
Description

A LoRA-based training and fine-tuning API for researchers and developers that exposes four functions (forward_backward, optim_step, sample, save_state) so users can run supervised learning and reinforcement learning experiments on a broad lineup of open-source models (Qwen, GPT-OSS, DeepSeek, Moonshot Kimi, NVIDIA Nemotron) without managing distributed compute infrastructure. Pricing is per million tokens with separate Prefill/Sample/Train rates and storage at $0.10/GB-month.

2TML-Interaction-Small
CategoryFoundation Model
Description

Thinking Machines Lab's proprietary multimodal interaction model (276B parameter MoE with 12B active) that natively handles real-time full-duplex audio, video, and text interaction using 200ms time-aligned micro-turns with encoder-free early fusion. Achieves 0.40s turn-taking latency and 77.8 on FD-bench v1.5, outperforming GPT-realtime-2.0 and Gemini-3.1-flash-live. Currently available as a limited research preview.

3Connectionism
CategoryResearch Publication
Description

Thinking Machines Lab's research blog named after the 1980s AI subfield. Publishes technical blog posts, papers, and code on topics including defeating nondeterminism in LLM inference, LoRA fine-tuning, on-policy distillation, and interaction models.

Scale indicator13 records

Each record includes

Type, Value, Description, Source

Partnership9 partners
Strategic tierFlagshipTypeStrategic or Co-development PartnerAnnounced on2026-04-22
Description

Multi-billion dollar agreement giving Thinking Machines Lab access to Google Cloud's A4X Max virtual machines with NVIDIA GB300 GPUs as one of the first customers on Blackwell architecture. Includes Google Kubernetes Engine, Spanner, Cluster Director, Cloud Storage, and Anywhere Cache. Early testing showed 2X improvement in training and serving speeds.

Strategic tierCoreTypeTechnology or IntegrationAnnounced on2026-04-22
Description

Google Cloud's Jupiter network for weight transfers in reinforcement learning workloads, providing enhanced training performance through network optimization for AI workloads.

Strategic tierCoreTypeStrategic or Co-development PartnerAnnounced on2026-03-20
Description

Co-members of the NVIDIA Nemotron Coalition advancing open AI through shared research, expertise, data, and compute on NVIDIA DGX Cloud infrastructure. First joint project is a base model co-developed with Mistral AI.

Strategic tierCoreTypeStrategic or Co-development PartnerAnnounced on2026-03-16
Description

Founding member of the NVIDIA Nemotron Coalition alongside Black Forest Labs, Cursor, LangChain, Mistral AI, Perplexity, Reflection AI, and Sarvam. Coalition will co-develop open frontier-level AI models trained on NVIDIA DGX Cloud infrastructure, with the first project being a base model co-developed with Mistral AI to underpin the Nemotron 4 family.

Strategic tierCoreTypeStrategic or Co-development PartnerAnnounced on2026-01-01
Description

$471.7 million non-cancelable contract for AI cloud infrastructure services; reported as a meaningful contract win for the BRUN side while TML provides compute/infrastructure services.

Strategic tierFlagshipTypeTechnology or Integration
Description

Listed as authorized subcontractor providing global cloud computing infrastructure for the Tinker service, as documented in the DPA Annex 2.

Strategic tierCoreTypeTechnology or Integration
Description

Authorized subcontractor providing Content Delivery Network (CDN) services globally for the Tinker platform as documented in DPA Annex 2.

Strategic tierCoreTypeTechnology or Integration
Description

Authorized subcontractor providing authentication services in the USA for the Tinker platform as documented in DPA Annex 2.

9Qwen (Alibaba)
Strategic tierCoreTypeTechnology or Integration
Description

Tinker supports Qwen family of models for fine-tuning, including Qwen3.6-35B-A3B, Qwen3.6-27B, Qwen3.5-4B, Qwen3.5-9B, and several other variants. On-policy distillation research used Qwen3-8B-Base with Qwen3-32B as teacher.

thinkingmachines.ai
Recent move8 records

Each record includes

Date, Type, Title, Description, Source

Expansion highlight6 records

Each record includes

Type, Description

Peers10 records
TypeEmerging player
Description

Managed fine-tuning and RL API for production LLM agents. Closest direct product analog to Tinker at the developer-platform layer for fine-tuning open models.

TypeDirect peer
Description

Enterprise-focused foundation model company offering customizable LLMs and fine-tuning APIs. Competes with TML on enterprise customization, though Cohere emphasizes closed models versus TML's open-model orientation.

TypeEmerging player
Description

Cloud platform for running and fine-tuning open-source ML models via API. Overlaps with Tinker's API-first, open-model fine-tuning positioning for developers.

TypeDirect peer
Description

Frontier AI lab building foundation models (GPT-4o, Realtime API) and developer APIs. Most direct competitor to TML's Interaction Model and Tinker fine-tuning API; TML was founded by former OpenAI executives and shares the same multimodal interaction and customization thesis.

TypeDirect peer
Description

Frontier AI lab with Claude models and an emphasis on AI safety and customization. Competes with TML on foundation model quality, developer tooling, and safety positioning, and has hired some former TML founders.

TypeDirect peer
Description

European open-weights model developer and Nemotron Coalition co-member alongside TML. Competes directly on open-weight foundation models and fine-tuning services that overlap with Tinker's value proposition.

TypeDirect peer
Description

Elon Musk's frontier AI lab (Grok) backed by massive compute. Competes for the same frontier-research talent pool that TML has lost employees to, and operates in the same multimodal/generative AI category.

TypeBroad incumbent
Description

Alphabet's frontier research lab producing Gemini and Gemini Live, which set the benchmarks TML's Interaction Model competes against. DeepMind is also the parent of Google Cloud, TML's multi-billion-dollar infrastructure partner.

TypeDirect peer
Description

Open-model cloud platform offering fine-tuning, inference, and training APIs over open-source LLMs. Most directly comparable product competitor to Tinker's fine-tuning-as-a-service offering.

TypeEmerging player
Description

Frontier open-weight model startup and Nemotron Coalition co-member. Competes with TML on open frontier model development and shares NVIDIA compute infrastructure.

Market position
Strengths5 records

Each record includes

Headline, Details, Source

Weaknesses5 records

Each record includes

Headline, Details, Source

Competitive moat4 records

Each record includes

Type, Details

Key risks7 records

Each record includes

Headline, Details, Source

Key highlights7 records

Each record includes

Headline, Details, Source

Customer concentration

Classification, Details

Named customers7 records

Each record includes

Name, Industry, Type, Use case, Source, UUID

Ideal customer profile3 records

Each record includes

Profile, Firmographic size, Sales motion, Sales cycle length, Buying structure, Purchase trigger, Buyer persona, Geography, Industry vertical, Primary use case, Description, Pain points, Evidence proof points, Target buyer

Technology focused
Yes
API detail
Has APIbool
Yes

Docs URL, Description

Integration9 records

Each record includes

Title, Type, Description, Source

AI capability13 records

Each record includes

Type, Description, Source

AI maturity
App detail

Has app

Feature9 records

Each record includes

Title, Differentiator, Description, Source

Core technology
Revenue estimate
Valuation estimate
Number of profiles
Profiles5 records

Each record includes

Name, Designation, Designation category, Overview, Profile commentary, Source

No data
Compliance4 records

Each record includes

Name, Class, Description

Funding overview

Funding stage, Last funding date, Total funding USD

Funding rounds3 records

Each record includes

Round, Amount USD, Date, Pre money valuation, Total investors, Investors, News

Investors17 records

Each record includes

Name, Type, Date of entry, Rounds participated, Website

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A1 record

Each record includes

Name, Acquisition type, Announced date, Completed date, Status, Website, News

Investment1 record

Each record includes

Name, Round, Announced date, Lead investor, Website, News

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Thinking Machines Lab

Foundation Models / AI Infrastructurethinkingmachines.ai

Thinking Machines Lab is a San Francisco-based AI research and product company building frontier multimodal interaction models (TML-Interaction-Small) and a fine-tuning API (Tinker), serving researchers, developers, and enterprises with customizable, real-time human-AI collaboration systems.

What Thinking Machines Lab does

Thinking Machines Lab is a private, AI research and product company founded in February 2025 in San Francisco by former OpenAI CTO Mira Murati alongside a group of OpenAI alumni co-founders. The company's stated mission is to make AI systems more widely understood, customizable, and generally capable, with an explicit emphasis on human-AI collaboration rather than fully autonomous agents. Its product surface centers on two pillars: Tinker, a LoRA-based training and fine-tuning API launched in October 2025 that abstracts distributed compute for researchers running RL and supervised learning experiments across a broad lineup of open-source models (Qwen, GPT-OSS, DeepSeek, Moonshot Kimi, NVIDIA Nemotron); and TML-Interaction-Small, a proprietary 276B-parameter mixture-of-experts multimodal model (12B active) released as a research preview in May 2026 that natively processes audio, video, and text in 200ms time-aligned micro-turns and achieves state-of-the-art FD-bench v1.5 and turn-taking latency scores.

The underlying technology stack combines proprietary model architecture (encoder-free early fusion, dMel audio embedding, batch-invariant inference kernels) with significant published research contributions on the Connectionism blog covering LoRA fine-tuning methodology, on-policy distillation (demonstrating 9–30x compute cost reductions), and deterministic inference. The company has locked in frontier compute capacity through a multi-year, gigawatt-scale NVIDIA Vera Rubin deployment (targeting early 2027) and a multi-billion-dollar Google Cloud agreement on GB300-powered A4X Max VMs, with NVIDIA also making a significant equity investment.

Thinking Machines Lab's commercial model is primarily API-first and usage-based, with Tinker priced per million tokens across Prefill, Sample, and Train operations plus $0.10/GB-month storage. Early customers skew academic and research-oriented (UC Berkeley, Princeton, Stanford, Redwood Research), with enterprise revenue anchored by a $471.7M non-cancelable cloud-infrastructure contract with Boost Run and adjacent infrastructure services supplied to customers like RadixArk. Distribution runs through self-serve developer signup supplemented by direct enterprise agreements and cloud marketplace channels; revenue remains undisclosed and the company has been capital-intensive rather than profit-oriented since founding, with a $2B seed round at a $12B valuation in July 2025 underwriting ongoing buildout.

Thinking Machines Lab firmographics

Firmographics
Name
Thinking Machines Lab
Legal name
Thinking Machines Lab, Inc.
Website
https://thinkingmachines.ai
Company type
Private
Founded year
2025
Operating status
Operating
Headcount range
101–250 employees
Short description
Thinking Machines Lab is a San Francisco-based AI research and product company building frontier multimodal interaction models (TML-Interaction-Small) and a fine-tuning API (Tinker), serving researchers, developers, and enterprises with customizable, real-time human-AI collaboration systems.
Ownership category
akta.pro rank

Thinking Machines Lab industry classification

Industry
Product category
Foundation Models / AI Infrastructure
NAICS
Scientific Research and Development Services (5417), Computer Systems Design and Related Services (54151), Other Computer Related Services (541519)
SIC
Services-Computer Programming Services (7371)
akta.pro industry
PRIMARYHDAAACAC
Fine-Tuning, Adaptation & Custom Model Training (PEFT/LoRA/RLHF)
Hardware, Technology & Telecom/Artificial Intelligence & Machine Learning/Foundation Models & Generative AI (LLMs, multimodal)
SECONDARYHDAEANAB
Model Development & Training Platforms (AutoML, Notebooks, Feature Stores)
Hardware, Technology & Telecom/Enterprise Platforms & Databases/Artificial Intelligence Platforms for Enterprise
SECONDARYHDAEANAD
LLMOps & Generative AI Platforms (Prompt/Agent Orchestration, RAG)
Hardware, Technology & Telecom/Enterprise Platforms & Databases/Artificial Intelligence Platforms for Enterprise
SECONDARYHDAAAAAI
AI Compiler, Runtime & Kernel Optimization Software (CUDA/ROCm/XLA, graph compilers)
Hardware, Technology & Telecom/Artificial Intelligence & Machine Learning/AI Infrastructure & Compute (GPUs/ASICs, accelerators)

Keywords

  • Multimodal AI models
  • AI fine-tuning API
  • Foundation model research
  • AI infrastructure services
  • Open-source AI training

Where Thinking Machines Lab is headquartered

Location

Headquarters

HQ city
San Francisco
HQ country
United States
HQ region
North America

Offices2 records

Markets served

Thinking Machines Lab business model

Business model
GTM type
B2B
Offering type
Software
Cost components
Infrastructure, Personnel, Technology or R&D, Marketing or Sales, Operations

Revenue model

  1. Tinker API Consumption: Usage-based revenue from token consumption (Prefill, Sample, Train pricing per million tokens across multiple model families including Qwen, GPT-OSS, DeepSeek, Kimi, NVIDIA Nemotron). Storage charged at $0.10/GB-month. Pre-paid credits system with promotional grants available.
  2. Enterprise Service Agreements: Pre-paid credits for Services (paid API/hosted model access); Enterprise Agreements possible for enterprise customers with proprietary terms; non-refundable fees with 1-year expiration on pre-paid credits
  3. Multi-Billion-Dollar Cloud/Compute Contracts: Multi-year infrastructure deals with Google Cloud (single-digit billions) and NVIDIA (gigawatt-scale Vera Rubin deployment) for AI model training and deployment capacity

Pricing tiers

ModelBillingPrice
Usage-basedPay-as-you-goNVIDIA Nemotron-3-Nano-30B-A3B-BF16 - small MoE tier at $0.13 Prefill / $0.33 Sample / $0.40 Train per million tokens (50% limited-time discount)
Usage-basedPay-as-you-goNVIDIA Nemotron-3-Super-120B-A12B-BF16 - mid-tier model at $0.38 Prefill / $0.96 Sample / $1.16 Train per million tokens (50% discount)
Usage-basedPay-as-you-goNVIDIA Nemotron-3-Ultra-550B-A55B-BF16 - largest model at $1.66-$3.32 Prefill / $4.15-$8.30 Sample / $4.98-$9.96 Train per million tokens (50% discount)
Usage-basedPay-as-you-goGPT-OSS-20B at $0.12 Prefill / $0.30 Sample / $0.36 Train per million tokens (32K context)
Usage-basedPay-as-you-goQwen3.5-397B-A17B - $2.00-$4.00 Prefill / $5.00-$10.00 Sample / $6.00-$12.00 Train per million tokens
Usage-basedPay-as-you-goKimi K2.5 - $1.47 Prefill / $3.66 Sample / $4.40 Train per million tokens; retiring July 12
Usage-basedPay-as-you-goStorage: $0.10 per GB-month across all tiers

Go-to-market motion1 record

Distribution channels3 records

Marketing channels8 records

Thinking Machines Lab product offering

Product offering

Core offering

Thinking Machines Lab builds and sells AI research products, primarily the Tinker fine-tuning API for researchers and developers (enabling LoRA-based training on open-source models like Qwen, GPT-OSS, DeepSeek, and NVIDIA Nemotron) and proprietary multimodal interaction models (TML-Interaction-Small) that natively handle real-time audio, video, and text collaboration with 200ms micro-turn latency. The company sells usage-based API access to its training platform and grants research preview access to its frontier interaction models.

Product overview

Thinking Machines Lab offers a platform-plus-modules product portfolio anchored on two main pillars: (1) Tinker, a managed fine-tuning API launched in October 2025 that lets researchers and developers apply LoRA-based supervised learning and RL training to a broad lineup of open-source models (Qwen, GPT-OSS, DeepSeek, Moonshot Kimi, NVIDIA Nemotron) without managing GPU infrastructure; and (2) TML-Interaction-Small, a proprietary 276B-parameter multimodal interaction model introduced as a research preview in May 2026 that natively handles full-duplex audio, video, and text interaction in 200ms micro-turns. The Connectionism research blog and News page round out the public-facing surface, publishing technical posts (LoRA Without Regret, On-Policy Distillation, Defeating Nondeterminism in LLM Inference), product announcements, and research grant programs (Interactivity Research Grants, Tinker Research and Teaching Grants) that bridge research outputs to the Tinker training platform and the company's frontier interaction models.

Differentiator

Problem solved

Functional benefit

Brands

  • Tinker: A training API for researchers providing fine-tuning of open-source models with LoRA; handles scheduling, tuning, resource management, and infrastructure reliability. Includes a Tinker cookbook and a public X account @tinkerapi.
  • Connectionism
  • TML-Interaction-Small

Products and services

  • Tinker A LoRA-based training and fine-tuning API for researchers and developers that exposes four functions (forward_backward, optim_step, sample, save_state) so users can run supervised learning and reinforcement learning experiments on a broad lineup of open-source models (Qwen, GPT-OSS, DeepSeek, Moonshot Kimi, NVIDIA Nemotron) without managing distributed compute infrastructure. Pricing is per million tokens with separate Prefill/Sample/Train rates and storage at $0.10/GB-month.
  • TML-Interaction-Small Thinking Machines Lab's proprietary multimodal interaction model (276B parameter MoE with 12B active) that natively handles real-time full-duplex audio, video, and text interaction using 200ms time-aligned micro-turns with encoder-free early fusion. Achieves 0.40s turn-taking latency and 77.8 on FD-bench v1.5, outperforming GPT-realtime-2.0 and Gemini-3.1-flash-live. Currently available as a limited research preview.
  • Connectionism Thinking Machines Lab's research blog named after the 1980s AI subfield. Publishes technical blog posts, papers, and code on topics including defeating nondeterminism in LLM inference, LoRA fine-tuning, on-policy distillation, and interaction models.

Companies that use Thinking Machines Lab

Customer profile

Named customers7 records

Ideal customer profiles3 records

Thinking Machines Lab technology and API

Technology

Technology focussed Yes

API detail

Has API
Yes
API docs
API detail

Core technology

AI maturity

App detail

Integration9 records

AI capability13 records

Feature9 records

Thinking Machines Lab partnerships and signals

Strategic signal

Partnerships

Nine partnerships are on record, tiered flagship and core.

  • Google CloudflagshipStrategic or Co-development Partner · 22 April 2026Multi-billion dollar agreement giving Thinking Machines Lab access to Google Cloud's A4X Max virtual machines with NVIDIA GB300 GPUs as one of the first customers on Blackwell architecture. Includes Google Kubernetes Engine, Spanner, Cluster Director, Cloud Storage, and Anywhere Cache. Early testing showed 2X improvement in training and serving speeds.
  • Jupiter Network (Google Cloud internal)coreTechnology or Integration · 22 April 2026Google Cloud's Jupiter network for weight transfers in reinforcement learning workloads, providing enhanced training performance through network optimization for AI workloads.
  • Black Forest Labs, Cursor, LangChain, Mistral AI, Perplexity, Reflection AI, SarvamcoreStrategic or Co-development Partner · 20 March 2026Co-members of the NVIDIA Nemotron Coalition advancing open AI through shared research, expertise, data, and compute on NVIDIA DGX Cloud infrastructure. First joint project is a base model co-developed with Mistral AI.
  • NVIDIA Nemotron CoalitioncoreStrategic or Co-development Partner · 16 March 2026Founding member of the NVIDIA Nemotron Coalition alongside Black Forest Labs, Cursor, LangChain, Mistral AI, Perplexity, Reflection AI, and Sarvam. Coalition will co-develop open frontier-level AI models trained on NVIDIA DGX Cloud infrastructure, with the first project being a base model co-developed with Mistral AI to underpin the Nemotron 4 family.
  • Boost Run Inc. (BRUN)coreStrategic or Co-development Partner · 1 January 2026$471.7 million non-cancelable contract for AI cloud infrastructure services; reported as a meaningful contract win for the BRUN side while TML provides compute/infrastructure services.
  • Google Cloud (Subcontractor)flagshipTechnology or IntegrationListed as authorized subcontractor providing global cloud computing infrastructure for the Tinker service, as documented in the DPA Annex 2.
  • CloudflarecoreTechnology or IntegrationAuthorized subcontractor providing Content Delivery Network (CDN) services globally for the Tinker platform as documented in DPA Annex 2.
  • WorkOScoreTechnology or IntegrationAuthorized subcontractor providing authentication services in the USA for the Tinker platform as documented in DPA Annex 2.
  • Qwen (Alibaba)coreTechnology or IntegrationTinker supports Qwen family of models for fine-tuning, including Qwen3.6-35B-A3B, Qwen3.6-27B, Qwen3.5-4B, Qwen3.5-9B, and several other variants. On-policy distillation research used Qwen3-8B-Base with Qwen3-32B as teacher.

Scale indicators13 records

Recent moves8 records

Expansion highlights6 records

Thinking Machines Lab competitors and assessment

Company assessment

Emerging players

  • OpenPipe: Managed fine-tuning and RL API for production LLM agents. Closest direct product analog to Tinker at the developer-platform layer for fine-tuning open models.
  • Replicate: Cloud platform for running and fine-tuning open-source ML models via API. Overlaps with Tinker's API-first, open-model fine-tuning positioning for developers.
  • Reflection AI: Frontier open-weight model startup and Nemotron Coalition co-member. Competes with TML on open frontier model development and shares NVIDIA compute infrastructure.

Direct peers

  • Cohere: Enterprise-focused foundation model company offering customizable LLMs and fine-tuning APIs. Competes with TML on enterprise customization, though Cohere emphasizes closed models versus TML's open-model orientation.
  • OpenAI: Frontier AI lab building foundation models (GPT-4o, Realtime API) and developer APIs. Most direct competitor to TML's Interaction Model and Tinker fine-tuning API; TML was founded by former OpenAI executives and shares the same multimodal interaction and customization thesis.
  • Anthropic: Frontier AI lab with Claude models and an emphasis on AI safety and customization. Competes with TML on foundation model quality, developer tooling, and safety positioning, and has hired some former TML founders.
  • Mistral AI: European open-weights model developer and Nemotron Coalition co-member alongside TML. Competes directly on open-weight foundation models and fine-tuning services that overlap with Tinker's value proposition.
  • xAI: Elon Musk's frontier AI lab (Grok) backed by massive compute. Competes for the same frontier-research talent pool that TML has lost employees to, and operates in the same multimodal/generative AI category.
  • Together AI: Open-model cloud platform offering fine-tuning, inference, and training APIs over open-source LLMs. Most directly comparable product competitor to Tinker's fine-tuning-as-a-service offering.

Broad incumbents

  • Google DeepMind: Alphabet's frontier research lab producing Gemini and Gemini Live, which set the benchmarks TML's Interaction Model competes against. DeepMind is also the parent of Google Cloud, TML's multi-billion-dollar infrastructure partner.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat4 records

Key risks7 records

Key highlights7 records

Customer concentration

Thinking Machines Lab social profiles

Digital presence

Thinking Machines Lab compliance and trust

Trust signal

Compliance4 records

Thinking Machines Lab financial estimates

Financial estimate

Revenue estimate

Valuation estimate

Thinking Machines Lab leadership team

Management profile

Number of profiles

Profiles5 records

Thinking Machines Lab funding detail

Funding detail

Funding overview

Funding rounds3 records

Investors17 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

Thinking Machines Lab M&A and investment

M&A and investment

M&A1 record

Investments1 record

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Thinking Machines Lab

What does Thinking Machines Lab do?

Thinking Machines Lab builds and sells AI research products, primarily the Tinker fine-tuning API for researchers and developers (enabling LoRA-based training on open-source models like Qwen, GPT-OSS, DeepSeek, and NVIDIA Nemotron) and proprietary multimodal interaction models (TML-Interaction-Small) that natively handle real-time audio, video, and text collaboration with 200ms micro-turn latency. The company sells usage-based API access to its training platform and grants research preview access to its frontier interaction models.

Is Thinking Machines Lab a public or private company?

Thinking Machines Lab is a private company. It is classified as venture growth investor backed and is currently operating.

When was Thinking Machines Lab founded?

Thinking Machines Lab was founded in 2025. It employs 101 to 250 people.

Where is Thinking Machines Lab based?

Thinking Machines Lab is headquartered in San Francisco, United States, in the North America region.

How does Thinking Machines Lab make money?

Three revenue lines are on record. Tinker API Consumption is the primary driver. The others are enterprise Service Agreements and multi-Billion-Dollar Cloud/Compute Contracts.

Who are Thinking Machines Lab's main competitors?

Emerging players on record are OpenPipe, Replicate and Reflection AI. Direct peers are Cohere, OpenAI, Anthropic, Mistral AI, xAI and Together AI. Google DeepMind is listed as a broad incumbent.

Does Thinking Machines Lab have an API?

Yes. Public-facing Tinker training API for researchers and developers to fine-tune open-source models using LoRA. Provides four core functions (forward_backward, optim_step, sample, save_state) that abstract away distributed training, GPU cluster scheduling, and resource management. Supports a broad lineup of open-source models including Qwen3.5/3.6 series (dense and MoE), GPT-OSS-120B/20B, DeepSeek-V3.1, Moonshot Kimi K2.5/K2.6, and NVIDIA Nemotron-3 Nano/Super/Ultra. Pricing is per million tokens (Prefill, Sample, Train), with storage charged at $0.10/GB-month. Users can sign up via auth.thinkingmachines.ai; universities and organizations can request wide-scale access by emailing tinker@thinkingmachines.ai. The API enables building custom fine-tuned models, RL training workflows, and on-policy distillation recipes (described in the Tinker cookbook). Developer documentation is at tinker-docs.thinkingmachines.ai.

What industry is Thinking Machines Lab in?

Thinking Machines Lab's product category is Foundation Models / AI Infrastructure. Its primary akta.pro industry code is HDAAACAC, Fine-Tuning, Adaptation & Custom Model Training (PEFT/LoRA/RLHF), with a secondary code of HDAEANAB, Model Development & Training Platforms (AutoML, Notebooks, Feature Stores). Its NAICS code is 5417 and its SIC code is 7371.

Unlock the full company data

50 free credits on sign-up, no credit card required.

Contact sales
Live signals
StartupHub.aiAI Research Interns Run Tests, Humans Keep TasteOpenAI and Thinking Machines researchers say AI research interns now run end-to-end experiments, but humans retain final judgment. Progress is about 80% on AGI and 25-50% on recursive self-improvement, with human taste still essential. Labs expect RL from researcher feedback to bake judgment into models.THE DECODERThinking Machines' Inkling 2 Small drops 316 F-bombs across benchmark runsResearchers at Vals AI found 316 instances of swear words in Inkling 2 Small's internal thinking logs across 2,655 runs. Most came from two runaway runs, including one with 158 expletives. The profanity does not signal failure, as 75% of scored runs fail anyway.WowtaleThinking Machines Lab Unveils Inkling 2 Small as It Reportedly Nears a $5B RoundThinking Machines Lab released Inkling 2 Small, a 594B-parameter model, and plans to open weights under Apache 2.0 after safety testing. The company is close to a $5B round led by Accel at a $40B pre-money valuation. The model ranks second on four of six coding benchmarks but trails top closed models.VentureBeatThinking Machines launches Inkling 2 Small with near-frontier performance on enterprise benchmarksThinking Machines launched Inkling 2 Small on November 5, a 594-billion-parameter model with 17 billion active per token. It scores 31 on Artificial Analysis and 71.8% on EnterpriseBench-CoreCraft, beating Mistral Large 4. Weights will be released under Apache 2.0 after safety testing.SiliconANGLEThinking Machines debuts Inkling 2 Small amid reported $5B fundraising push - SiliconANGLEThinking Machines Labs debuted Inkling 2 Small, a large language model that automates AI training tasks, on Thursday. The company is reportedly raising $5 billion led by Accel, with a $40 billion valuation. The model includes a vision encoder and is available via Tinker cloud, with weights to be released under an open-source license next week.WebProNewsThinking Machines Puts Inkling 2 Small on Tinker, With Apache 2.0 Weights Planned Next WeekThinking Machines launched Inkling 2 Small, a 594B MoE model with 17B active parameters, on its Tinker platform. The model scored 31 on the Artificial Analysis Intelligence Index, up from 13 for its predecessor, and accepts image inputs. Apache 2.0 weights are planned next week after safety testing.THE DECODERThinking Machines' Inkling 2 Small claims top spot among western open-weight modelsThinking Machines released Inkling 2 Small, a 594B-parameter mixture-of-experts model with 17B active parameters, scoring 31 on the Artificial Analysis Intelligence Index. It costs $1.18 per task, six times its predecessor, and is available via Tinker with Apache 2.0 weights planned.Dealroom.coThinking Machines Lab nears $5B round at $40B valuation, led by AccelThinking Machines Lab is raising $5 billion at a $40 billion pre-money valuation, led by Accel, with the round expected to close within days. The company has raised about $7 billion in disclosed funding, and its Tinker system and Inkling model are key products. The raise more than triples its prior $12 billion valuation.Tech Funding NewsReport: Accel is to lead a $5B for Thinking Machines Lab at a $40B valuation — TFNAccel is leading a $5 billion round at a $40 billion pre-money valuation for Thinking Machines Lab, potentially closing in the next few days. The valuation is over three times the seed round and about 20% lower than the 2025 figure, with annual revenue exceeding $100 million. The company plans to release full model weights under an Apache 2.0 license after safety testing.Trending TopicsInkling 2: Thinking Machines Overtakes Mistral in Open Weights | Trending TopicsThinking Machines unveiled Inkling 2 Small, a 594-billion-parameter model that scores 31 points on Artificial Analysis, beating Mistral Large 4's 26. The company is seeking a $5 billion funding round at a $40 billion valuation. Weights are expected to be released under Apache 2.0 after safety testing.