Identifai Ltd Announces Research Publication & Corporate Update

THE INFORMATION CONTAINED HEREIN IS RESTRICTED AND IS NOT FOR PUBLICATION, RELEASE OR DISTRIBUTION, DIRECTLY OR INDIRECTLY, IN THE UNITED STATES OF AMERICA OR ANY OTHER JURISDICTION IN WHICH SUCH PUBLICATION, RELEASE OR DISTRIBUTION WOULD BE UNLAWFUL.

LONDON, UK / ACCESS Newswire / October 1, 2026 / Research Publication & Corporate Update

Identifai Ltd, a UK-based artificial intelligence company building intelligent automation for the financial services sector, today announces the publication of its first two research papers and provides an update on its fourth funding round. In August, the Company published its benchmark results. Today, it publishes the full research behind them - the methodology, the data, and the evaluation infrastructure - and places all of it on the public record for independent scrutiny.

Key Points

• The Company has published its first research paper, Model Audit: Benchmarking LLM Agents on Financial Spreadsheets. Model Audit is a benchmark built to answer a question no existing evaluation could: how well do AI agents actually perform the spreadsheet work of professional finance? The benchmark comprises 131 complex tasks and 3,225 expert-vetted evaluation criteria, constructed with 78 finance professionals across seven countries - drawn from investment banking, private equity, hedge funds and asset management - with manipulation tasks representing at least 45 minutes of analyst work each. The findings are stark: every frontier model evaluated, including GPT-5.5, Claude Sonnet 4.6, Gemini 3.1 Pro and Grok 4.20, scores below 50% on average. Identifai 1.0 is among the strongest models evaluated, leads all models on synthesis tasks at 66.7%, and posts the highest overall score on the publicly released task set. In the accompanying human utility study, finance professionals rated Identifai 1.0's outputs as a usable starting point more often than any other model's. The full evaluation harness has been open-sourced for independent verification.
• The Company has published its second research paper, Spreadsheet Wars: Decomposing Preference in LLM Generation of Spreadsheet Workbooks. The paper introduces the Spreadsheet Wars platform: blind, head-to-head evaluation of AI-generated spreadsheets, decided by user preference votes. Across 4,357 blind votes spanning 16 frontier models, Identifai 1.0 ranked first - with an Elo rating of 1550 against GPT-4o's baseline of 1000, ahead of Claude Sonnet 4.5, Claude Opus 4.1 and Gemini 3 Pro - and retained first place after statistical adjustment for surface features. Identifai 1.0 recorded the highest win rate on the platform at 83.5%. In the paper's expert evaluation, conducted by professionals with investment banking and private equity backgrounds, Identifai 1.0 was the only model to consistently follow professional financial modelling standards.
• The research is public, and deliberately so. Both papers, including methodology, data and results, are available in full from the Company. A research function producing peer-grade published work is rare at Identifai's stage; the Company regards it as fundamental. Claims made to the market should be checkable, and as of today, Identifai's are.
• The Spreadsheet Wars platform now underpins the Company's public launch event. The live Spreadsheet Wars event, confirmed for February 2027, will see first-year analysts from leading investment banks compete alongside frontier AI models on the platform introduced in the paper - turning the Company's published research into a public stage.
• The fourth funding round opens on 1 October 2026 at £4.82 per Ordinary A share. Following the closing of the third round at £3.12, the new price represents a 382% cumulative increase from the inaugural seed price of £1.00. The Company is now in final discussions with Haatch Fund, an established EIS fund specialising in high-quality software businesses, regarding its intention to take up the fourth round in full, with completion expected during October 2026. An institutional, diligence-led fund taking an entire round at £4.82 per share would represent the strongest external validation in the Company's history - and it arrives in the same announcement as the research that supports it.

Commenting on today's update, George Croitor, CEO of Identifai, said:

"In August we published our numbers. Today we publish the work behind them. Anyone - investor, institution or competitor - can now read exactly how our results were produced, run the open-source harness, and check them independently. Two papers, one conclusion: the general-purpose models were not built for the work finance actually runs on, and ours was. That is why analysts preferred our spreadsheets in over eight out of ten blind contests, and it is why an institutional fund that has done its diligence is in final discussions to take the entire fourth round at £4.82. The research is published, the round opens today, and February's launch is getting closer with every announcement."

Enquiries:

Identifai
Emily Richardson
+44 (0) 20 8178 8417
[email protected]

Disclaimer: This announcement is for information purposes only and does not constitute or form part of an offer or invitation to acquire or dispose of any securities of Identifai or constitute a solicitation of any offer to purchase or subscribe for securities in any jurisdiction. All subscription dates are subject to change without notice.

This information is provided by Reach, the non-regulatory press release distribution service of RNS, part of the London Stock Exchange. Terms and conditions relating to the use and distribution of this information may apply. For further information, please contact [email protected] or visit www.rns.com.

SOURCE: Identifai Ltd