Skip to content
EN
Log inBook a demoOpen your account

Intelligence Layer

One account, one ledger, one agent layer. The modules are what you switch on.

Platform overview
The AI Intelligence Layer

One intelligence layer under every module: the same data, the same agents, the same ledger, whichever function you switch on.

  • One layer, every module
  • Shared data and agents
  • Every run on one ledger
Go to The AI Intelligence Layer
Isometric stack of pale grey slabs wired together on white, with one layer filled deep cyanStart hereThe architecture is the argumentWhy adding the sixth module costs a fraction of buying the first.Read the architecture
The Architecture

One account carries every brand you run. Modules are switched on per brand, and every run lands in the same ledger.

  • One account, many brands
  • Modules toggle per brand
  • A single billing ledger
Go to The Architecture
Isometric stack of pale grey slabs wired together on white, with one layer filled deep cyanStart hereThe architecture is the argumentWhy adding the sixth module costs a fraction of buying the first.Read the architecture
The agent layer

Three kinds of agent. Watchers notice, drafters produce, and Ask Intelligence answers from your own numbers.

  • Watchers run on a schedule
  • Drafters wait for sign-off
  • Answers carry their source
Go to The agent layer
Isometric stack of pale grey slabs wired together on white, with one layer filled deep cyanStart hereThe architecture is the argumentWhy adding the sixth module costs a fraction of buying the first.Read the architecture
Security

Isolated tenancy, roles that mean something, and a plain answer about what leaves your account.

  • Per-tenant isolation
  • Roles down to the module
  • Stated data boundaries
Go to Security
Isometric stack of pale grey slabs wired together on white, with one layer filled deep cyanStart hereThe architecture is the argumentWhy adding the sixth module costs a fraction of buying the first.Read the architecture

Seats are free — you pay for what the AI runs

Book a demoOpen your account

By role

The same product, read by job. For the person who holds the job, and for the company where the platform is the hire.

All roles
CEO / Director

Five dashboards, the ones you describe in a sentence, a weekly review that explains what moved, and a question box that quotes the real figure.

  • A weekly review with causes
  • Dashboards from a sentence
  • Caps on every dollar spent
Go to CEO / Director
Marketing

Brand Content, Social, Search, and Ads on one account, with seven watchers that report what moved and a voice you teach once.

  • 37 text formats, priced runs
  • Six networks, real limits
  • Google ranks and AI answers
Go to Marketing
Sales

A pipeline that won’t let a deal sit without a next step and a date, red dots on the quiet ones, and follow-up mails drafted in your own Gmail.

  • Overdue and quiet deals flagged
  • Drafts wait for your Send
  • Two currencies on every amount
Go to Sales
Project Manager

Kanban and Gantt over the same cards, projects only their team can see, and work that arrives from findings, deals, posts, and validation requests.

  • Five ways a card arrives
  • A team and a planned cost
  • The Friday report as a PDF
Go to Project Manager
Customer Service

An assistant on your website that answers from what you gave it, says so when it doesn’t know, and logs every question it couldn’t answer.

  • On any site, one pasted line
  • Refuses to guess
  • A daily budget per desk
Go to Customer Service
Finance

The pipeline as numbers, every amount in two currencies, and a ledger and caps on every AI run.

  • The pipeline, weighted by stage
  • Two currencies, one rate table
  • Every run on the ledger
Go to Finance
Web

Rank tracking and the keyword gap, page formats that read the page they replace, and a desk on the site that sets no cookie.

  • Ranks in 232 markets
  • Pages written to a format
  • A desk on the site, one line
Go to Web
Human Resources

No HR module, said first. Drafting in the Playground, a handbook employees can question, and onboarding as tasks with owners and dates.

  • A handbook that answers
  • Onboarding as dated tasks
  • Free accounts for every hire
Go to Human Resources

Seats are free — you pay for what the AI runs

Book a demoOpen your account

Services

The people around the product, from the first connection to a named operator on your account.

All services
Onboarding

At least four hours, live: your account set up with you, your team trained, then three months of support.

  • 1,000 USD, 300 back in your wallet
  • At least four hours, live
  • Three months of support
Go to Onboarding
A fanned stack of white ruled sheets with the top one edged in deep cyanLarger organizationsRoll out by brand, by team, by functionOne account, one ledger, roles down to the function. Staged so procurement never has to open a file.See the group rollout
Data integration

We connect the systems you already run and prepare the data behind them, so every answer has a source.

  • Connectors built and tested
  • Metrics defined once
  • Sources reconciled
Go to Data integration
A fanned stack of white ruled sheets with the top one edged in deep cyanLarger organizationsRoll out by brand, by team, by functionOne account, one ledger, roles down to the function. Staged so procurement never has to open a file.See the group rollout
Custom features

An agent or a function built for your process, on the same ledger and under the same approval gates.

  • Scoped before it is priced
  • Same gates, same ledger
  • Yours to keep
Go to Custom features
A fanned stack of white ruled sheets with the top one edged in deep cyanLarger organizationsRoll out by brand, by team, by functionOne account, one ledger, roles down to the function. Staged so procurement never has to open a file.See the group rollout
Dedicated hosting

A single-tenant deployment in the region you choose, with the data boundary written down and an uptime commitment.

  • Single tenant
  • Region of your choice
  • Uptime in writing
Go to Dedicated hosting
A fanned stack of white ruled sheets with the top one edged in deep cyanLarger organizationsRoll out by brand, by team, by functionOne account, one ledger, roles down to the function. Staged so procurement never has to open a file.See the group rollout
Premium Support

A named contact, committed response times, and a quarterly review of what ran, what it cost and what to change.

  • Named contact
  • Committed response times
  • Quarterly review
Go to Premium Support
A fanned stack of white ruled sheets with the top one edged in deep cyanLarger organizationsRoll out by brand, by team, by functionOne account, one ledger, roles down to the function. Staged so procurement never has to open a file.See the group rollout
Training

Sessions for the people who will approve, review and ask, so the workforce is used rather than watched.

  • By role, not by feature
  • Live on your account
  • Recorded for the next hire
Go to Training
A fanned stack of white ruled sheets with the top one edged in deep cyanLarger organizationsRoll out by brand, by team, by functionOne account, one ledger, roles down to the function. Staged so procurement never has to open a file.See the group rollout
Managed operations

Someone reviews what the watchers found, approves the drafts within your caps, and runs the weekly review.

  • Drafts approved in your name
  • Caps respected
  • Weekly review delivered
Go to Managed operations
A fanned stack of white ruled sheets with the top one edged in deep cyanLarger organizationsRoll out by brand, by team, by functionOne account, one ledger, roles down to the function. Staged so procurement never has to open a file.See the group rollout

Seats are free — you pay for what the AI runs

Book a demoOpen your account

About

Why this was built, who is behind it, what it has done for the companies running it, and how to run it well.

About BearingBridge
Why we built it?

What a company can do has been capped by who it could afford to hire. That cap is the thing that moved.

  • Capability, not headcount
  • Written, not benchmarked
  • No lock-in claim
Go to Why we built it?
A row of pale grey blocks climbing steeply to a single tall deep cyan block, a hairline curve tracing the riseThe short versionPunch above your headcountThe AI workforce for companies that need more capability than they can staff.Read the argument
Who is behind BearingBridge?

Who built this, what it is independent of, and why that independence is worth stating out loud.

  • Model-independent
  • No data resale
  • Named people behind it
Go to Who is behind BearingBridge?
A row of pale grey blocks climbing steeply to a single tall deep cyan block, a hairline curve tracing the riseThe short versionPunch above your headcountThe AI workforce for companies that need more capability than they can staff.Read the argument
How we run the work

The six-phase method every account is built on: a written bearing, a dated baseline, a pilot on real data, and a fix that reconciles the money.

  • A kill criterion, in writing
  • Costs watched as they run
  • An ending without us
Go to How we run the work
A row of pale grey blocks climbing steeply to a single tall deep cyan block, a hairline curve tracing the riseThe short versionPunch above your headcountThe AI workforce for companies that need more capability than they can staff.Read the argument
Case studies

Ten builds told the way they ran: the situation, what was built, what changed months later, and what got switched off.

  • Sector and function stated
  • Client-verified figures
  • The stopped work left in
Go to Case studies
A row of pale grey blocks climbing steeply to a single tall deep cyan block, a hairline curve tracing the riseThe short versionPunch above your headcountThe AI workforce for companies that need more capability than they can staff.Read the argument
FAQ

The questions that come up before every demo, answered here so the demo can be about your business.

  • Product and billing
  • Security and data
  • Answered in plain words
Go to FAQ
A row of pale grey blocks climbing steeply to a single tall deep cyan block, a hairline curve tracing the riseThe short versionPunch above your headcountThe AI workforce for companies that need more capability than they can staff.Read the argument
Testimonials

Customers on the module they actually run, quoted directly, with the module named.

  • Module named each time
  • Quoted, not paraphrased
  • Role and size given
Go to Testimonials
A row of pale grey blocks climbing steeply to a single tall deep cyan block, a hairline curve tracing the riseThe short versionPunch above your headcountThe AI workforce for companies that need more capability than they can staff.Read the argument
Changelog

What shipped, dated, newest first. The place to check whether the thing you were promised exists yet.

  • Dated entries
  • Shipped only
  • Linked to the module
Go to Changelog
A row of pale grey blocks climbing steeply to a single tall deep cyan block, a hairline curve tracing the riseThe short versionPunch above your headcountThe AI workforce for companies that need more capability than they can staff.Read the argument
Partner program

Run the platform for the companies you advise. Every client is a brand on your account, billed on its own ledger.

  • Clients as brands
  • Separate ledgers
  • Margin on every run
Go to Partner program
A row of pale grey blocks climbing steeply to a single tall deep cyan block, a hairline curve tracing the riseThe short versionPunch above your headcountThe AI workforce for companies that need more capability than they can staff.Read the argument

Seats are free — you pay for what the AI runs

Book a demoOpen your account

Pricing

Seats are free. You pay for what the AI actually runs, and you see the price first.

Full pricing
How it works

Three sentences, and that is the whole model. No tiers to decode, no per-seat arithmetic to do.

  • No seat licence
  • No annual lock-in
  • Three sentences long
Go to How it works
Ten pale meter tracks filled to different levels in deep cyan, cut by one horizontal cyan limit lineNo surprisesYou see the price before you run itThe price, the caps and the gates, all visible before anything spends.Open pricing
Seats

Give an account to everyone who needs one. The number on the invoice does not move when you do.

  • Unlimited accounts
  • Zero per-seat cost
  • Roles still enforced
Go to Seats
Ten pale meter tracks filled to different levels in deep cyan, cut by one horizontal cyan limit lineNo surprisesYou see the price before you run itThe price, the caps and the gates, all visible before anything spends.Open pricing
What things cost

Image and video show their price before you run them, everything else its cost the moment it ends, and the machine’s own mistakes are not billed to you.

  • A price on every run
  • Retries are on us
  • Itemized in the ledger
Go to What things cost
Ten pale meter tracks filled to different levels in deep cyan, cut by one horizontal cyan limit lineNo surprisesYou see the price before you run itThe price, the caps and the gates, all visible before anything spends.Open pricing
Caps and gates

A daily allowance per person, a monthly cap for the organization and a budget per agent, and an approval gate in front of anything that publishes or spends.

  • Caps per person and agent
  • Approval before publish
  • A zero balance stops the AI
Go to Caps and gates
Ten pale meter tracks filled to different levels in deep cyan, cut by one horizontal cyan limit lineNo surprisesYou see the price before you run itThe price, the caps and the gates, all visible before anything spends.Open pricing
Enterprise

Hosting in Europe, the Americas or Asia, your own model keys, your company’s skills kept current, single sign-on and an annual contract, on quote.

  • Your own model keys
  • Single sign-on and SCIM
  • Annual contract, on quote
Go to Enterprise
Ten pale meter tracks filled to different levels in deep cyan, cut by one horizontal cyan limit lineNo surprisesYou see the price before you run itThe price, the caps and the gates, all visible before anything spends.Open pricing
Questions

The questions people actually ask about the bill, answered on the page rather than in a call.

  • Overage answered
  • Cancellation answered
  • Migration answered
Go to Questions
Ten pale meter tracks filled to different levels in deep cyan, cut by one horizontal cyan limit lineNo surprisesYou see the price before you run itThe price, the caps and the gates, all visible before anything spends.Open pricing

Seats are free — you pay for what the AI runs

Book a demoOpen your account

Insights

Three collections, one standard: long-form, sourced, dated, and never behind a form.

All insights
AI Trends

What is moving under the industry: model economics, the Chinese price tier, and how buyers now ask assistants instead of searching.

  • The model market, read closely
  • AI search and citations
  • Every claim sourced and dated
Go to AI Trends
A white leaderboard sheet of six rows on a pale grid, its top row edged and filled in deep cyan, the others in greyModel benchmarksEvery model, scored in publicSourced scores for the text, image and video models in the app, refreshed twice a month. Each figure links to the page it was read on.Compare the models
Best Practices — AI Guide

The working guide: when to buy and when to build, when an agent is the wrong tool, and what survives contact with production.

  • Build-or-buy, decided
  • Architectures that ship
  • Prompts that do real work
Go to Best Practices — AI Guide
A white leaderboard sheet of six rows on a pale grid, its top row edged and filled in deep cyan, the others in greyModel benchmarksEvery model, scored in publicSourced scores for the text, image and video models in the app, refreshed twice a month. Each figure links to the page it was read on.Compare the models
CEO's Opinion

Signed columns from the person running the company: what building agents across six functions actually shows, ahead of the industry line.

  • Signed, never ghostwritten
  • From live builds, not decks
  • Positions, not press releases
Go to CEO's Opinion
A white leaderboard sheet of six rows on a pale grid, its top row edged and filled in deep cyan, the others in greyModel benchmarksEvery model, scored in publicSourced scores for the text, image and video models in the app, refreshed twice a month. Each figure links to the page it was read on.Compare the models
AI model benchmarks

Every model you can pick in the app, for text, image and video, scored on public benchmarks, with the model behind Quick, Balanced and Best.

  • Sortable scores, each one sourced
  • A price level, never an amount
  • No figure ever estimated
Go to AI model benchmarks
A white leaderboard sheet of six rows on a pale grid, its top row edged and filled in deep cyan, the others in greyModel benchmarksEvery model, scored in publicSourced scores for the text, image and video models in the app, refreshed twice a month. Each figure links to the page it was read on.Compare the models

Seats are free — you pay for what the AI runs

Book a demoOpen your account

Insights/AI Trends/AI in practice

What an open-weight AI model is, and why it matters.

DeepSeek, Kimi and GLM now sit beside Claude and GPT in most model lists, ours included. What sets them apart is one published file. Here is what that file is, and what it changes for a buyer.

71.5%

of tokens on Vercel's AI Gateway went to open-weight models in September 2026

22.6%

of the estimated spend on that gateway in the same month

4 months

average lag of the best open-weight model behind the closed frontier, per Epoch AI

8

open-weight text models in BearingBridge's pickers on November 1, 2026

BearingBridgeNovember 20266 min read

A brass clock movement lying open on a worn wooden workbench, its gears visible, with loose gears, fine tools and a loupe on a linen cloth and a closed metal case out of focus behind
On this sheet

Open a model list today and you’ll see two kinds of names. Claude, GPT and Gemini on one side. DeepSeek, Kimi and GLM on the other. The second group is usually called open weight. The term gets used a lot and explained rarely, so this page explains it.

The weights are the model

A language model is a very long list of numbers. Training sets those numbers, called weights or parameters, and they hold everything the model learned. Kimi K3.1 Pro, one of the largest, has 2.8 trillion of them.

With a closed model, the maker keeps that file. You send a question to the maker’s servers and get an answer back. You rent access and never hold the model.

With an open-weight model, the maker publishes the file. Anyone can download it and run it on their own machines, or pay a hosting company to run it for them.

Open weight isn’t open source

The two terms get mixed up. The Open Source Initiative, which keeps the definition of open source software, wrote one for AI too. Its Open Source AI Definition asks for the freedom to use, study, modify and share a system. That takes three things: the weights, the code that trained the model, and detailed information about the training data.

Open weight covers the first of the three. The OSI’s own page on the term says it refers to the final weights of a trained network, without the training code or the data. You can run the model and adapt it. You can’t rebuild it or see what it learned from.

The license varies too. DeepSeek publishes under MIT, which allows commercial use. Kimi K3, GLM-5.3 and Meta’s Muse Spark Open each come with their maker’s own license, which Artificial Analysis sums up as “commercial use is allowed with restrictions.” Read the license before you build a product on one.

What a published file changes

More than one company can serve it. A closed model is sold by its maker and a few cloud partners. On November 20, Artificial Analysis counted 24 API providers for GLM-5.3 and 21 for DeepSeek V4.1 Flash. Hosts compete on price and speed for the same model. That is a large part of why the cheap tier is cheap, which we covered in the model mix piece.

You choose where it runs. A bank can run the file in its own data center, and nothing leaves the building. A smaller firm can pick a host in the country its contracts require. With a closed model, the maker decides where the servers are.

It can’t be taken back. A host can drop a model from its catalogue. A published file stays published, and another host can pick it up.

It can be adapted. A team with the skills can retrain the file on its own documents. Few small firms will, but the vendors they buy from do.

Buyers have noticed. On Vercel’s AI Gateway, a service many software teams use to reach models, open-weight models carried 71.5% of tokens in September 2026. In December 2025 it was 7%. Vercel notes that its definition of open weight is broader than in its earlier reports, so the two figures aren’t strictly like for like. The direction isn’t in doubt.

Are they as good?

They’re close, and how close depends on who measures.

Epoch AI, a research group, found on May 29 that since January 2026 the best open-weight models have trailed the best closed ones by four months on average. SemiAnalysis, a research firm, went further on August 21. In the free part of its study, it writes that “with each generation, open-source models take half as long to catch up.” In the current generation, by its count, Kimi K2.6 passed Claude Opus 4.5 on its test suite 4.8 months after Opus came out.

Then the tests got harder to study for. On October 27, Artificial Analysis rebuilt its Intelligence Index so that 60% of it rests on private test sets. The gap between the best open-weight model and the best closed one went from 3 points to 12. Kimi K3.1 Pro scored 46, and Claude Opus 5.5 scored 58.

SemiAnalysis adds a caution of its own. A score isn’t the whole story, and its analysts say they still prefer a closed model for their daily work.

So on public tests the two groups are within a few points of each other, and on tests nobody has seen the closed models keep a lead. For drafting, summarizing, sorting and translating, that lead rarely shows. On long, hard jobs it still can.

What you give up

Running the file yourself takes hardware and people. Nobody runs 2.8 trillion parameters on an office laptop. Most firms that use open-weight models rent them from a host, and then the host’s terms apply, not only the license.

Open doesn’t always mean cheap, either. In the same Vercel report, DeepSeek V4.1 Flash handled nearly half of all September tokens for 2.6% of the spend. Kimi K3 was 3.2% of the tokens and drove 47.6% of the rise in open-weight spend. That’s how open-weight models as a group reached 22.6% of spend.

And you still don’t know what the model was trained on.

How we use them at BearingBridge

We offer both kinds. On November 1, our model benchmarks page listed eight open-weight text models from five makers: DeepSeek V4 Pro, V4 Flash and V4.1 Flash, Kimi K3 and K3.1 Pro, GLM 5.3, Muse Spark Open and Mistral Large 4. We checked each one’s status on its Artificial Analysis page on November 20. They sit in the same pickers as Claude, GPT and Gemini.

We don’t run the weights on our own servers. Every call, to an open-weight model or a closed one, goes through one gateway to a hosted endpoint, and lands in the same ledger with its cost.

Three things follow for a customer:

  • Your admin decides. Each feature has its own list of allowed models. An organization that wants no open-weight model, or none from a given maker, switches them off.
  • You can compare on your own work. The Playground sends one question to up to three models side by side and prints the price under each answer. That tells you more than a leaderboard does.
  • The price level is on the page. The two DeepSeek Flash models sit at the lowest of our four price levels. Kimi K3 and K3.1 Pro sit at the highest, beside the most expensive closed models.

We refresh the scores on that page twice a month. The habit we’d suggest to anyone: open-weight models for the bulk work, a closed model where a wrong answer costs money, and a test on real tasks before either becomes the default.

Try it this week

Pick a job your team runs every day, like a product description or a summary of a customer call. Send it to one closed model and one open-weight model. Have the person who knows the work grade both without seeing which is which.

Then ask your vendors which models sit behind their features, and who hosts them.

Last reviewed November 20, 2026. Every source above was opened that day. Model lists, licenses and scores change often. Corrections to hello@bearingbridge.com.

Start

Reading about it is the slow way.

Open an account and run one module on one brand, or book a demo and see the platform run live.

No subscription. No seat fee. No card required to open an account.