AI for management · updated 8 September 2026
Which AI can a management team trust?
- Part of the AI Board research programme
Memory: the advertised window
The largest published context window went from 4.1K to 10M tokens, so document length is rarely the limit any more.
| Point | Value |
|---|---|
| ChatGPT (GPT-3.5) | 4.1K |
| GPT-4 | 8.19K |
| Claude (100K) | 100K |
| GPT-4 Turbo | 128K |
| Claude 2.1 | 200K |
| Gemini 1.5 Pro (2M) | 2M |
| Llama 4 Scout | 10M |
Practical memory versus advertised
The advertised window is not working memory: one of the strongest models in the NoLiMa study lost 29.6 percentage points of accuracy by 32,000 tokens.
| Label | Value |
|---|---|
| Short text | 99.3% |
| At 32,000 tokens | 69.7% |
Confident mistakes per generation
Between GPT-4o and GPT-5 in thinking mode, the share of answers carrying a major factual error fell by a factor of 4.3.
| Label | Value |
|---|---|
| GPT-4o (May 2024) | 20.6% |
| o3 (Apr 2025) | 22.0% |
| GPT-5 main (Aug 2025) | 11.6% |
| GPT-5 thinking (Aug 2025) | 4.8% |
EU residency among flagships
9 of the 13 current flagship models have a documented route to running with data residency inside the EU.
- Documented EU route9
- No documented EU route4
| Segment | Value | Share |
|---|---|---|
| Documented EU route | 9 | 69% |
| No documented EU route | 4 | 31% |
Cost per executive per working week
At public list prices, one executive's week of heavy use costs between US$0.89 and US$11.88, before any discount.
| Label | Value |
|---|---|
| GPT-6 Astra | US$11.88 |
| Claude Fable 5.1 | US$10.62 |
| Claude Opus 5 | US$5.94 |
| Qwen3.8-Max | US$5.16 |
| Mistral Medium 3.5 | US$4.05 |
| Grok 4.6 | US$2.64 |
| GPT-5.6 Terra | US$2.50 |
| Claude Sonnet 5 | US$2.38 |
| DeepSeek-V4-Pro | US$1.26 |
| Gemini 3.8 Flash | US$0.89 |
What benchmarks do not answer
Five questions a management team actually asks, and no public benchmark answers any of them on your own board pack.
| Label | Value |
|---|---|
| Memory | 4 |
| Factuality | 3 |
| Honesty | 2 |
| Citation | 2 |
| Discipline | 3 |
02Landscape now
The current flagship models, side by side
Every current flagship in the dataset, with the facts a management team decides on: how much it can read, whether it can run in the EU, and what it costs per million tokens.
| Lab | Model | Memory | EU residency | In / M | Out / M |
|---|---|---|---|---|---|
| OpenAI | GPT-6 Astra | 1.05M tokens | YesEU region | US$10.00 | US$50.00 |
| Anthropic | Claude Fable 5.1 | 1M tokens | YesEU multi-region | US$10.00 | US$50.00 |
| Anthropic | Claude Mythos 5.1 | 1M tokens | No | US$10.00 | US$50.00 |
| Google (DeepMind) | Gemini 4 Argon | 1.05M tokens | YesEU multi-region | US$2.00 | US$10.00 |
| xAI | Grok 4.7 | 500K tokens | No | US$2.00 | US$6.00 |
| Meta | Muse Wonder | 1M tokens | No | US$4.00 | US$20.00 |
| DeepSeek | DeepSeek-V4-Pro | 1M tokens | Yesself-hosted | US$1.32 | US$3.96 |
| Mistral AI | Mistral Medium 3.5 | 256K tokens | YesEU region | US$1.50 | US$7.50 |
| Alibaba (Qwen) | Qwen3.8-Max | 1M tokens | YesEU region | US$2.00 | US$6.00 |
| Z.ai | GLM-5.3 | 1M tokens | No | US$1.40 | US$4.40 |
| Moonshot AI | Kimi K3 | 1.05M tokens | Yesself-hosted | US$3.00 | US$15.00 |
| Cohere | Command A+ | 128K tokens | Yesself-hosted | not published | not published |
| NVIDIA | Nemotron 3 Ultra | 1M tokens | Yesself-hosted | not published | not published |
See the full landscape, including the tier below and the models being retired
03The research behind the numbers
Three pages, one question
Nobody reads a research report end to end. The charts above are the summary; these pages are where each number is argued, sourced and dated.
AI Board Model Index, first edition
The ranking lands on this page. There are no results yet and no scores anywhere on this site, because nothing has been measured. When the first edition is published, the leaderboard opens here and the method page becomes its appendix.
04For your role
What this means at your seat
The same landscape looks different from three chairs. Each role has a section in the guide and a page on what an AI assistant does in that job.
CEO
Synthesis across a messy pile of documents, and an answer you can repeat in a meeting without checking it first.
CFO
Citation discipline. A number without a document and a page number is an opinion, however confident it sounds.
CTO
Residency, licences and the exit route. The generation you can host in Europe is not always the newest one.
05Method
How we work
This page follows the same discipline as the rest of the site: nothing is published that cannot be checked.
- Primary sources only
- Every fact about a model comes from vendor documentation, an official launch post, a system card, a peer-reviewed paper, or a cloud provider's own region list. Never from a summary of a summary.
- Facts as of a stated date
- Model specifications change weekly. Every chart prints the date its figures were verified, so you can see how old a claim is before you act on it.
- What we cannot verify, we leave out
- If a figure has no primary source, it does not appear on the page, not as an estimate and not as a range. A shorter page is the price of a checkable one.
- No vendor money
- No model vendor pays for a place in this research, and none of them sees a page before it is published.
- Corrections
- Found something that is wrong or out of date? Tell us and we correct the page, with the change dated. Contact details are on the about page.
06Every seat at the table
AI per management role
The research is built around three chairs: CEO, CFO and CTO. Every other seat at the management table has its own page on what an AI assistant grounded in your own files does in that job, each at its own altitude. AI for CMOs is about defending the marketing number in front of the CFO; the page for marketing managers is about producing it. The question most managers ask first, will AI replace managers, has its own sourced answer.
Leadership and strategy
- AI for CEOs
Walk into the MT with numbers you can defend
- AI for general managers
The P&L is yours, the staff functions aren't
- AI for strategy managers
Know whether the plan is landing, before the MT asks
Finance
- AI for finance leaders
Ask your own numbers, get the answer before the meeting
- AI for controllers
Answers you can re-check, grounded in your own numbers
Operations, procurement and service
- AI for COOs
One ops number that reconciles
- AI for operations managers
Delivery & capacity
- AI for procurement managers
OTIF, price increases and contract renewals, answered from your own files
- AI for customer service managers
The AI that answers to you, not to your customers
HR
- AI for HR directors
Workforce cost, attrition and succession, grounded in your own files
- AI for HR managers
Absence, retention and headcount, answered from your own files
IT and engineering
- AI for CTOs
Ship dates, cloud cost, pentest triage and tech debt, answered from your own files
- AI for CIOs
Your application landscape, security posture and shadow IT, answered from your own files
- AI for IT managers
Your own IT data, answered before the CFO, the auditor, or the renewal deadline forces it
- AI for engineering managers
Walk in with the cause named, not the apology
Product
- AI for product leaders
Portfolio decisions grounded in your own product data
- AI for product managers
Stop guessing whether anyone uses what you shipped
Marketing and sales
- AI for CMOs
The marketing number that survives the CFO opening the model
- AI for marketing managers
Prove pipeline contribution and where every euro went
- AI for sales managers
Pipeline truth, forecast honesty, and the deals about to slip
Legal and compliance
- AI for in-house counsel
Know what is true in your own registers
Become AI-native before your competition does
Ride the AI wave instead of swimming behind it. Book a demo and we'll schedule your install.
Your company brain lives on your laptop. You choose if a question goes to a cloud model or stays fully local.