Fewer Tokens, Better Answers: Give Your Agent Search, Not Raw Files Even with just 295 pages, agents benefit from search. We ran the same Claude agent on the same corpus two ways: 📂 Using its native tools to explore the corpus, inspect documents, and decide what to read 🔍 Using LightOn API for Search On a 5,000-page knowledge base, LightOn delivered: ✅ 36% fewer tokens ✅ Better answers in 75% of blind evaluations ✅ Flat search costs as the corpus grew 17× The interesting part isn't that long context windows eventually hit their limits. It's that search already improves efficiency and answer quality long before context becomes the bottleneck. A larger context window gives your agent more room, but it doesn't magically make it better at retrieval. Full breakdown 👇🏻 lighton.ai/lighton-blogs/few… Start saving tokens now with : console.lighton.ai

لأن اللغة العربية تستحق أن تُفهم كما كُتبت Arabic pushes OCR systems in ways few languages do LightOnOCR-2 has been adapted to Arabic through targeted fine-tuning, demonstrating the model's ability to handle right-to-left text, connected cursive script, and complex document layouts Across the Middle East and North Africa, organizations manage vast volumes of Arabic-language documents in environments where data control matters as much as performance For these organizations, data sovereignty and language coverage are not competing priorities. Both are essential Because document AI shouldn't stop where language complexity begins Read the full breakdown : lighton.ai/lighton-blogs/lig…

Performance *and* sovereingty: Make your knowledge agent-readable. Production agents need a layer that can parse messy documents, structure knowledge, retrieve the right evidence, and cite its sources. LightOn Console is that knowledge layer. One API. Three endpoints. Free tier.

Open Source LLMs do extremely well on very difficult tasks like BrowseCompPlus thanks to the LightOn Open Source technology (Reason-ModernColbert, LateOn, Agent-ModernColbert,...) Available to developers now!

Open Source LLMs do extremely on very difficult tasks like BrowseCompPlus thanks to the LightOn Open Source technology (Reason-ModernColbert, LateOn, Agent-ModernColbert,...) Available to developers now!

Short reminder: Open Source LLMs can do extremely well compared to proprietary models on very difficult tasks like BrowseCompPlus thanks to the Open Source technology developed by @LightOnIO (Reason-ModernColbert, LateOn, Agent-ModernColbert,...) that is available to developers at console.lighton.ai

Today, we're introducing LightOn Console. ⚙️ Three endpoints: /Parse any documents /Extract structured data /Search enterprise knowledge with citations 🔌 Built-in connectors. MCP-ready. Governance enforced at the chunk level. No infrastructure. No pipeline maintenance. No dedicated retrieval team required. Make your enterprise knowledge agent-readable now! Read the launch announcement: lighton.ai/lighton-blogs/int… Test it now: console.lighton.ai/

🔍 Confidence is good. Citations are mandatory. An agent can produce a confident answer about a contract, a policy, or a regulation. It can sound right. It can also be wrong. Users do not need eloquence. They need the clause, the source, and the page. With LightOn Console, /search returns answers grounded in your enterprise knowledge base, with citations to the relevant chunks. That is the difference between sounding right and being verifiable. 📄 Test it tomorrow, June 2, 2026. 👉🏻 Join the waiting list. 🔗 console.lighton.ai

50 million downloads on @huggingface! LightOn SOTA late-interaction and dense retrievers, OCR models, and LLMs are validated by the community and tested every day in production. 🧪 LightOn is now one of the most active labs in the world in retrieval, pushing the Pareto frontier of AI, always open source. 🙌 Huge kudos to @iacopo_poli, @AmelieTabatta, @antoine_chaffin, @raphaelsrty, @staghado, @CavaillesAdrien, @baptaubertin, Maxence Lasbordes, @paulomouraj and the entire team for the work! 🔥 As proud as LightOn is of this milestone, the model is only one building block. Orchestration is what makes AI systems work. Now it’s your turn to test them in production: console.lighton.ai/

LightOn annonce la signature d'un contrat pluriannuel avec INFOCOM'94, opérateur public numérique pour accélérer le déploiement de l’IA générative souveraine auprès de 26 collectivités d'Île-de-France. 💬 « Infocom'94 héberge les données applicatives de ses membres dans des domaines aussi sensibles que les finances, les ressources humaines, le scolaire, le logement social ou les CCAS. Ces données publiques citoyennes ne doivent pas sortir du périmètre concerné. LightOn était structurellement le mieux positionné pour répondre à cette contrainte. » Vincent Kerbiquet, Directeur d'Infocom'94 LightOn sera déployé en mode hybride : stockage des données sur site, puissance de calcul déportée, aucune rétention des données. 🔗 Lire le communiqué : lighton.ai/fr-blog-posts/lig…