Developer docs
API playgroundTry for free, no card

Search company profiles

Deep Cogito

Full company profile

uuid0000b9v

Namestring
Deep Cogito
Legal namestring
Deep Cogito Inc.
Websiteurl
deepcogito.com
Company typeenum
Private
Founded yearint
2025
Operating statusenum
Operating
Ownership categoryenum
Headcount rangeband
11–50
akta.pro rankint
HeadquartersSan Francisco, United States
HQ citystring
San Francisco
HQ countrystring
United States
HQ regionstring
North America
Markets served

Serves global market

Keyword5 values
open-weight language models, foundation model development, AI inference APIs, reasoning model training, open source AI
Industry2 codes
PRIMARYHDAAACAC
Fine-Tuning, Adaptation & Custom Model Training (PEFT/LoRA/RLHF)
Hardware, Technology & Telecom/Artificial Intelligence & Machine Learning/Foundation Models & Generative AI (LLMs, multimodal)
SECONDARYHDAAACAD
Prompt Engineering, Orchestration & LLMOps Tooling
Hardware, Technology & Telecom/Artificial Intelligence & Machine Learning/Foundation Models & Generative AI (LLMs, multimodal)
NAICS code2 codes
  • Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services518
  • Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services5182
SIC code1 code
  • Services-Prepackaged Software7372
Product category
Foundation Models
Social media profiles2 records
GTM motion2 records

Each record includes

Type, Description, Source

Revenue model1 record
1API Access / Model Inference
TypeUsage Based
Description

The model is available as an API on multiple cloud platforms including OpenRouter, Fireworks AI, Together AI, Ollama's cloud, Baseten, and RunPod, suggesting usage-based or subscription-based API access as a primary revenue stream.

deepcogito.com
Marketing channels3 records

Each record includes

Title, Type, Stage, Description, Source

Distribution channels4 records

Each record includes

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

Cost components5 values
Technology or R&D, Personnel, Infrastructure, Marketing or Sales, Operations
GTM typeB2B
B2B
Offering typeSoftware
Software
Core offering1 text field

Deep Cogito develops and releases open-weight large language models (LLMs) trained via process supervision for reasoning chains, with its flagship Cogito v2.1 671B model positioned as the best open-weight LLM from a US company. The company distributes the model weights openly on Hugging Face, provides API inference through multiple cloud platforms (OpenRouter, Fireworks AI, Together AI, Ollama Cloud, Baseten, RunPod), supports local deployment via Ollama and Unsloth, and offers a free web-based chat interface at chat.deepcogito.com.

Differentiator
Functional benefit
Problem solved
Quantifiable outcome1 value
  • Lowest average token usage among reasoning models of similar capabilities
Product overview1 text field

Deep Cogito offers a single unified product portfolio centered on the Cogito v2.1 671B large language model. The company provides the model through multiple channels: a free web-based chat interface (chat.deepcogito.com) for direct user access, and API access via multiple cloud platforms (OpenRouter, Fireworks AI, Together AI, Ollama Cloud, Baseten, RunPod) as well as local deployment options (Ollama, Unsloth). The model is trained via process supervision and achieves competitive benchmark performance while using fewer tokens than comparable reasoning models.

Product and service4 records
1Cogito v2.1 671B
CategoryFoundation Model
Description

Flagship 671 billion parameter open-weight large language model (MoE) trained via process supervision for reasoning chains, designed for general reasoning, instruction following, coding, longer queries, multi-turn conversations, and creativity tasks. Targeted at developers and AI researchers building applications that require cost-efficient frontier-quality reasoning.

2chat.deepcogito.com
CategoryAI Application
Description

Free web-based chat interface allowing users to interact directly with the Cogito v2.1 671B model. The service states it does not store chats and does not use inputs to train models, providing a direct-to-user channel to evaluate the model's reasoning and conversational capabilities.

3Cogito v2 Preview Models
CategoryFoundation Model
Description

Preview release of hybrid reasoning models in four sizes (70B dense, 109B MoE, 405B dense, 671B MoE), introducing hybrid reasoning and serving as the precursor to the v2.1 release. Available as open weights.

4Cogito v1 Model Family
CategoryFoundation Model
Description

Initial family of open-source models (3B, 8B, 14B, 32B, 70B parameters) trained using Iterated Distillation and Amplification (IDA) as a scalable alignment and self-improvement strategy, released when the company emerged from stealth.

Scale indicator2 records

Each record includes

Type, Value, Description, Source

Partnership10 partners
Strategic tierCoreTypeTechnology or Integration
Description

DeepCogito forks the open-licensed DeepSeek base model from November 2024 and runs post-training in-house. This forms the foundational technology base for the Cogito v2.1 model.

Strategic tierCoreTypeChannel Partner/ Reseller/ Distributor
Description

OpenRouter distributes Cogito v2.1 as an API endpoint, making the model accessible to developers through its platform.

Strategic tierCoreTypeChannel Partner/ Reseller/ Distributor
Description

Fireworks AI provides API access to Cogito v2.1 through its inference platform.

Strategic tierCoreTypeChannel Partner/ Reseller/ Distributor
Description

Together AI hosts and provides API access to Cogito v2.1 as part of its model offerings.

Strategic tierCoreTypeChannel Partner/ Reseller/ Distributor
Description

Ollama provides both cloud API access and local inference support for Cogito v2.1, enabling developers to run the model in the cloud or locally.

Strategic tierCoreTypeChannel Partner/ Reseller/ Distributor
Description

Baseten hosts Cogito v2.1 in its model library and provides API-based access to the model.

Strategic tierCoreTypeChannel Partner/ Reseller/ Distributor
Description

RunPod provides a playground endpoint for Cogito v2.1 671B FP8 dynamic through its console.

Strategic tierCoreTypeChannel Partner/ Reseller/ Distributor
Description

Hugging Face hosts the Cogito v2.1 model weights on its platform, making them freely available for download and serving via Hugging Face Inference Endpoints.

Strategic tierCoreTypeChannel Partner/ Reseller/ Distributor
Description

Unsloth provides local inference support for Cogito v2.1 in GGUF format, enabling efficient local model serving without cloud dependency.

10LM Studio
TypeChannel Partner/ Reseller/ Distributor
Description

LM Studio is listed as a supporting project alongside Unsloth and Ollama for enabling local inference of Cogito models.

Recent move6 records

Each record includes

Date, Type, Title, Description, Source

Expansion highlight6 records

Each record includes

Type, Description

Peers10 records
TypeDirect peer
Description

European open-weight LLM provider releasing competitive base and instruct models under permissive licenses. Most direct competitor to Deep Cogito in the open-weight reasoning space, with similar distribution through cloud APIs and Hugging Face.

TypeDirect peer
Description

Chinese AI lab whose open-licensed base model (Nov 2024) Deep Cogito explicitly forks. Directly comparable on open-weight MoE reasoning models (DeepSeek R1, V3) and benchmark positioning.

TypeBroad incumbent
Description

Meta's Llama family is the largest open-weight LLM ecosystem globally. Competes with Deep Cogito for the same developer and enterprise open-weight mindshare, though Meta is a much larger incumbent with vastly more compute.

4Alibaba (Qwen)
TypeDirect peer
Description

Qwen is a leading open-weight LLM family from Alibaba with strong reasoning models. Directly comparable in open-weight positioning, multilingual benchmarks, and developer adoption on Hugging Face.

TypeBroad incumbent
Description

Closed frontier model provider (GPT-4o, o1/o3) that Deep Cogito benchmarks against. Sets the upper bound on capability expectations and dominates enterprise API spend.

TypeBroad incumbent
Description

Closed frontier model provider (Claude family) competing for the same enterprise and developer workload as Deep Cogito's open-weight models, particularly in reasoning and coding.

TypeBroad incumbent
Description

Frontier closed model provider (Gemini) competing across reasoning, coding, and multimodal benchmarks that Deep Cogito targets with Cogito v2.1.

TypeEmerging player
Description

Enterprise-focused LLM provider offering both proprietary and open-weight models. Comparable as a horizontal LLM competitor, though more enterprise- and retrieval-focused than Deep Cogito.

TypeEmerging player
Description

Foundation model lab offering the Jamba family of open-weight hybrid SSM-Transformer models. Comparable as a smaller open-weight model provider pursuing enterprise and developer adoption.

TypeOthers
Description

Cloud inference and fine-tuning platform that both hosts Deep Cogito's models as a distribution partner and competes indirectly by offering its own fine-tuned open-weight model variants to enterprise customers.

Market position
Strengths5 records

Each record includes

Headline, Details, Source

Weaknesses5 records

Each record includes

Headline, Details, Source

Competitive moat5 records

Each record includes

Type, Details

Key risks5 records

Each record includes

Headline, Details, Source

Key highlights7 records

Each record includes

Headline, Details, Source

Customer concentration

Classification, Details

Segment1 record

Each record includes

Title, Type, Primary, Description, Pain point addressed, Use case, Source

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

Integration8 records

Each record includes

Title, Type, Description, Source

AI capability6 records

Each record includes

Type, Description, Source

AI maturity
App detail

Has app

Feature3 records

Each record includes

Title, Differentiator, Description, Source

Core technology
Revenue estimate
Valuation estimate
Number of profiles
Profiles2 records

Each record includes

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

No data
No data
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

Investors6 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&A

Each record includes

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

Investment

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 →

Deep Cogito

Foundation Modelsdeepcogito.com

Deep Cogito firmographics

Firmographics
Name
Deep Cogito
Legal name
Deep Cogito Inc.
Website
https://deepcogito.com
Company type
Private
Founded year
2025
Operating status
Operating
Headcount range
11–50 employees
Ownership category
akta.pro rank

Deep Cogito industry classification

Industry
Product category
Foundation Models
NAICS
Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518), Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (5182)
SIC
Services-Prepackaged Software (7372)
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)
SECONDARYHDAAACAD
Prompt Engineering, Orchestration & LLMOps Tooling
Hardware, Technology & Telecom/Artificial Intelligence & Machine Learning/Foundation Models & Generative AI (LLMs, multimodal)

Keywords

  • Open-weight language models
  • Foundation model development
  • AI inference APIs
  • Reasoning model training
  • Open source AI

Where Deep Cogito is headquartered

Location

Headquarters

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

Markets served

Deep Cogito business model

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

Revenue model

  1. API Access / Model Inference: The model is available as an API on multiple cloud platforms including OpenRouter, Fireworks AI, Together AI, Ollama's cloud, Baseten, and RunPod, suggesting usage-based or subscription-based API access as a primary revenue stream.

Go-to-market motion2 records

Distribution channels4 records

Marketing channels3 records

Deep Cogito product offering

Product offering

Core offering

Deep Cogito develops and releases open-weight large language models (LLMs) trained via process supervision for reasoning chains, with its flagship Cogito v2.1 671B model positioned as the best open-weight LLM from a US company. The company distributes the model weights openly on Hugging Face, provides API inference through multiple cloud platforms (OpenRouter, Fireworks AI, Together AI, Ollama Cloud, Baseten, RunPod), supports local deployment via Ollama and Unsloth, and offers a free web-based chat interface at chat.deepcogito.com.

Product overview

Deep Cogito offers a single unified product portfolio centered on the Cogito v2.1 671B large language model. The company provides the model through multiple channels: a free web-based chat interface (chat.deepcogito.com) for direct user access, and API access via multiple cloud platforms (OpenRouter, Fireworks AI, Together AI, Ollama Cloud, Baseten, RunPod) as well as local deployment options (Ollama, Unsloth). The model is trained via process supervision and achieves competitive benchmark performance while using fewer tokens than comparable reasoning models.

Differentiator

Problem solved

Functional benefit

Products and services

  • Cogito v2.1 671B Flagship 671 billion parameter open-weight large language model (MoE) trained via process supervision for reasoning chains, designed for general reasoning, instruction following, coding, longer queries, multi-turn conversations, and creativity tasks. Targeted at developers and AI researchers building applications that require cost-efficient frontier-quality reasoning.
  • chat.deepcogito.com Free web-based chat interface allowing users to interact directly with the Cogito v2.1 671B model. The service states it does not store chats and does not use inputs to train models, providing a direct-to-user channel to evaluate the model's reasoning and conversational capabilities.
  • Cogito v2 Preview Models Preview release of hybrid reasoning models in four sizes (70B dense, 109B MoE, 405B dense, 671B MoE), introducing hybrid reasoning and serving as the precursor to the v2.1 release. Available as open weights.
  • Cogito v1 Model Family Initial family of open-source models (3B, 8B, 14B, 32B, 70B parameters) trained using Iterated Distillation and Amplification (IDA) as a scalable alignment and self-improvement strategy, released when the company emerged from stealth.

Quantifiable outcome

  • Lowest average token usage among reasoning models of similar capabilities

Companies that use Deep Cogito

Customer profile

Segments1 record

Ideal customer profiles3 records

Deep Cogito technology and API

Technology

Technology focussed Yes

API detail

Has API
Yes
API docs
API detail

Core technology

AI maturity

App detail

Integration8 records

AI capability6 records

Feature3 records

Deep Cogito partnerships and signals

Strategic signal

Partnerships

Ten partnerships are on record, tiered core.

  • DeepSeekcoreTechnology or IntegrationDeepCogito forks the open-licensed DeepSeek base model from November 2024 and runs post-training in-house. This forms the foundational technology base for the Cogito v2.1 model.
  • OpenRoutercoreChannel Partner/ Reseller/ DistributorOpenRouter distributes Cogito v2.1 as an API endpoint, making the model accessible to developers through its platform.
  • Fireworks AIcoreChannel Partner/ Reseller/ DistributorFireworks AI provides API access to Cogito v2.1 through its inference platform.
  • Together AIcoreChannel Partner/ Reseller/ DistributorTogether AI hosts and provides API access to Cogito v2.1 as part of its model offerings.
  • OllamacoreChannel Partner/ Reseller/ DistributorOllama provides both cloud API access and local inference support for Cogito v2.1, enabling developers to run the model in the cloud or locally.
  • BasetencoreChannel Partner/ Reseller/ DistributorBaseten hosts Cogito v2.1 in its model library and provides API-based access to the model.
  • RunPodcoreChannel Partner/ Reseller/ DistributorRunPod provides a playground endpoint for Cogito v2.1 671B FP8 dynamic through its console.
  • Hugging FacecoreChannel Partner/ Reseller/ DistributorHugging Face hosts the Cogito v2.1 model weights on its platform, making them freely available for download and serving via Hugging Face Inference Endpoints.
  • UnslothcoreChannel Partner/ Reseller/ DistributorUnsloth provides local inference support for Cogito v2.1 in GGUF format, enabling efficient local model serving without cloud dependency.
  • LM StudioChannel Partner/ Reseller/ DistributorLM Studio is listed as a supporting project alongside Unsloth and Ollama for enabling local inference of Cogito models.

Scale indicators2 records

Recent moves6 records

Expansion highlights6 records

Deep Cogito competitors and assessment

Company assessment

Direct peers

  • Mistral AI: European open-weight LLM provider releasing competitive base and instruct models under permissive licenses. Most direct competitor to Deep Cogito in the open-weight reasoning space, with similar distribution through cloud APIs and Hugging Face.
  • DeepSeek: Chinese AI lab whose open-licensed base model (Nov 2024) Deep Cogito explicitly forks. Directly comparable on open-weight MoE reasoning models (DeepSeek R1, V3) and benchmark positioning.
  • Alibaba (Qwen): Qwen is a leading open-weight LLM family from Alibaba with strong reasoning models. Directly comparable in open-weight positioning, multilingual benchmarks, and developer adoption on Hugging Face.

Broad incumbents

  • Meta (Llama): Meta's Llama family is the largest open-weight LLM ecosystem globally. Competes with Deep Cogito for the same developer and enterprise open-weight mindshare, though Meta is a much larger incumbent with vastly more compute.
  • OpenAI: Closed frontier model provider (GPT-4o, o1/o3) that Deep Cogito benchmarks against. Sets the upper bound on capability expectations and dominates enterprise API spend.
  • Anthropic: Closed frontier model provider (Claude family) competing for the same enterprise and developer workload as Deep Cogito's open-weight models, particularly in reasoning and coding.
  • Google DeepMind (Gemini): Frontier closed model provider (Gemini) competing across reasoning, coding, and multimodal benchmarks that Deep Cogito targets with Cogito v2.1.

Emerging players

  • Cohere: Enterprise-focused LLM provider offering both proprietary and open-weight models. Comparable as a horizontal LLM competitor, though more enterprise- and retrieval-focused than Deep Cogito.
  • AI21 Labs: Foundation model lab offering the Jamba family of open-weight hybrid SSM-Transformer models. Comparable as a smaller open-weight model provider pursuing enterprise and developer adoption.

Others

  • Together AI: Cloud inference and fine-tuning platform that both hosts Deep Cogito's models as a distribution partner and competes indirectly by offering its own fine-tuned open-weight model variants to enterprise customers.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks5 records

Key highlights7 records

Customer concentration

Deep Cogito social profiles

Digital presence

Deep Cogito financial estimates

Financial estimate

Revenue estimate

Valuation estimate

Deep Cogito leadership team

Management profile

Number of profiles

Profiles2 records

Deep Cogito funding detail

Funding detail

Funding overview

Funding rounds3 records

Investors6 records

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

Deep Cogito M&A and investment

M&A and investment

M&A

Investments

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

Frequently asked questions about Deep Cogito

What does Deep Cogito do?

Deep Cogito develops and releases open-weight large language models (LLMs) trained via process supervision for reasoning chains, with its flagship Cogito v2.1 671B model positioned as the best open-weight LLM from a US company. The company distributes the model weights openly on Hugging Face, provides API inference through multiple cloud platforms (OpenRouter, Fireworks AI, Together AI, Ollama Cloud, Baseten, RunPod), supports local deployment via Ollama and Unsloth, and offers a free web-based chat interface at chat.deepcogito.com.

Is Deep Cogito a public or private company?

Deep Cogito is a private company. It is classified as venture growth investor backed and is currently operating.

When was Deep Cogito founded?

Deep Cogito was founded in 2025. It employs 11 to 50 people.

Where is Deep Cogito based?

Deep Cogito is headquartered in San Francisco, United States, in the North America region.

How does Deep Cogito make money?

One revenue line is on record: API Access / Model Inference.

Who are Deep Cogito's main competitors?

Direct peers on record are Mistral AI, DeepSeek and Alibaba (Qwen). Broad incumbents are Meta (Llama), OpenAI, Anthropic and Google DeepMind (Gemini). Emerging players are Cohere and AI21 Labs. Together AI is listed as an others.

Does Deep Cogito have an API?

Yes. Cogito v2.1 671B is available as an API through multiple platforms: OpenRouter, Fireworks AI, Together AI, Ollama's cloud, Baseten, and RunPod. The model can also be run locally using Ollama or Unsloth (GGUF format).

What industry is Deep Cogito in?

Deep Cogito's product category is Foundation Models. Its primary akta.pro industry code is HDAAACAC, Fine-Tuning, Adaptation & Custom Model Training (PEFT/LoRA/RLHF), with a secondary code of HDAAACAD, Prompt Engineering, Orchestration & LLMOps Tooling. Its NAICS code is 518 and its SIC code is 7372.

Unlock the full company data

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

Contact sales
Live signals
VC News DailyDeep Cogito Receives $43M Series A RoundDeep Cogito, a post-training research lab, announced a $43 million Series A led by TQ Ventures, with participation from Benchmark, Nexus Venture Partners, Atreides Management, South Park Commons, and Zscaler. The round brings total funding to over $56 million. The company builds frontier open-weight models and applies its post-training engine to create specialized models trained on enterprise data.WebrazziYapay Zeka Gündemi #60This week's AI news covered new products from Perplexity, Google, Alibaba, and Z.ai, alongside rising AI-driven cyberattacks and AGI discussions. Nvidia acquired Hugging Face for $12.9 billion, and Anthropic signed a $45 billion infrastructure deal with Nscale. OpenAI's Sam Altman said an AGI model could be ready by year-end.VentureburnDeep Cogito Raises $43 Million To Build AI Models That Improve ThemselvesSan Francisco AI startup Deep Cogito announced a $43 million Series A on August 26, 2026, led by TQ Ventures with participation from Benchmark, Nexus Venture Partners, Atreides Management, South Park Commons and Zscaler, bringing total funding over $56 million. The company's post-training techniques, including Iterated Distillation and Amplification, aim to make models smarter without new human data.citybizDeep Cogito Raises $43M to Scale Enterprise AI Model TrainingSan Francisco-based Deep Cogito raised $43 million in a Series A round led by New York's TQ Ventures, with participation from Benchmark, Nexus Venture Partners, Atreides Management, South Park Commons and Zscaler. The round brings total capital above $56 million, to fund model-training infrastructure and software for adapting open-weight AI models to proprietary data.The SaaS NewsDeep Cogito Raises $43M Series ASan Francisco-based Deep Cogito announced a $43 million Series A round led by TQ Ventures on August 26, 2026, with additional investors including Benchmark, Nexus Venture Partners, Atreides Management, South Park Commons and Zscaler. The round brings total funding to over $56 million, to be used for research and engineering teams, model-training infrastructure, future releases and enterprise work.NewsnReleasesDeep Cogito raises $43 million Series A for AI self-improvement researchSan Francisco-based Deep Cogito announced a $43 million Series A round led by TQ Ventures on Wednesday, with participation from Benchmark, Nexus Venture Partners, Atreides Management, South Park Commons and Zscaler, which joined as both customer and investor. The round brings total funding to over $56 million. The company plans to expand its research and engineering teams, scale training infrastructure and advance future Cogito model releases.WebrazziYapay zeka girişimi Deep Cogito, 43 milyon dolar yatırım aldıDeep Cogito raised $43 million in a Series A led by TQ Ventures, with participation from Benchmark, Nexus Venture Partners, and others. The 2024-founded company's total funding now exceeds $56 million. It will use the funds to expand its research team, grow corporate business, and develop new open-source AI models.FinSMEsDeep Cogito Raises $43M in Series A FundingDeep Cogito raised $43M in Series A funding led by TQ Ventures, with participation from Benchmark, Nexus Venture Partners, Atreides Management, South Park Commons, and Zscaler. The round brings total funding to over $56M. The company will use the funds to expand its research and engineering team, scale infrastructure for training frontier models, and advance future Cogito releases.MarketScreenerDeep Cogito Inc. announced that it has received $43 million in funding from TQ Ventures LLC, Atreides Management, LP, Benchmark, Nexus Venture Partners, Southpark Commons LLC, Zscaler, Inc.Deep Cogito Inc. closed a financing transaction on August 26, 2026, receiving $3,000,044 in its third and final tranche led by new investor TQ Ventures LLC. The round included participation from Atreides Management, LP, Benchmark, Nexus Venture Partners, Southpark Commons LLC and Zscaler, Inc., totaling $43,000,000.Unite.AIDeep Cogito Raises $43M Series A to Build the Post-Training Engine for Self-Improving AISan Francisco AI lab Deep Cogito raised a $43 million Series A led by TQ Ventures, with Benchmark, Nexus Venture Partners, Atreides Management, South Park Commons and Zscaler participating, bringing total funding to over $56 million. The company's Iterated Distillation and Amplification approach aims to improve models through post-training rather than larger pre-training runs, with its 671B Cogito v2.1 model producing reasoning chains roughly 60% shorter than DeepSeek R1.