Skip to content
World clockEU--:--UK--:--USA--:--CN--:--PLDEFRIT中文EN

portal about AI and technologyevents · analysis · interviews · technical background

Search
LIVE
AnalysisTinubu tells Akwa Ibom rally naira would be worthless without 2023 reforms07.11AnalysisJapan spent a record ¥15.3993tn buying yen from 30 July to 26 August07.11NewsMachado's return starts an early campaign in Venezuela with no election date set07.11AnalysisUnitree's Wang says GPT-6 robots paint without training07.11ExplainerICC requires Duterte to attend 30 November trial in person07.11NewsGoogle clarifies schools can keep Search while disabling Gemini apps07.11AnalysisBrightline parent entities file prearranged Chapter 11 with $490m for train operator07.11NewsRocket Lab reports eight of 18 SDA satellites in production07.11ExplainerPakistan unveils 2026-31 auto policy with five-year tariff cuts and EV tax breaks07.11NewsMHRA extends Casgevy licence to children aged five and over07.11
›

CERT-In gets access to Anthropic's unreleased Mythos model for cyber defence testing

India's CERT-In has been given access to Mythos, an unreleased Anthropic model, for cyber defence testing, according to reporting on the arrangement.

Grace OkonkwoGrace OkonkwoAI, models and technology
AI & modelsExplainerGrace OkonkwoPublished: 5 October 20263 min readSources 12
CERT-In gets access to Anthropic's unreleased Mythos model for cyber defence testing

CERT-In is India's national computer emergency response team. It has been given access to Mythos, an Anthropic model that has not been released publicly. The arrangement is for cyber defence testing, according to reporting on the deal.

The access matters because a government security body gets an unreleased model before general release. What that changes for defenders, and for Anthropic's release plans, is not yet clear. Neither side has said how long the access lasts or what testing it covers.

Mythos is described as unreleased. The public cannot use it, and outside researchers cannot inspect it. CERT-In handles computer security incidents in India. Testing a model in that setting is different from a commercial rollout. The reporting does not say whether other agencies have similar access. It also does not say whether the testing feeds back into the model's development.

"The arrangement gives CERT-In access to Mythos for cyber defence testing," the reporting says.

Anthropic has not announced a public release date for Mythos. The terms of the access, including any restrictions on how the model is used, have not been disclosed.

Comments 0

Sources

12
  1. 01Wave of affordable electric cars is boosting consumers' choice: sales of sub-€25,000 models set to rise sevenfoldEN
  2. 02Introducing Beam: Reflection's 501B open-weight modelEN
  3. 03Introducing d1: The most capable decision model, now with visionEN
  4. 04Interfaze-1-lite: the first open-weight model for deterministic taskEN
  5. 05Precogly (The open-source alternative to commercial threat modeling platforms)EN
  6. 06Study: Path discovered to make AI models red-flag their doubtful answersEN
  7. 07Jeff-Code: a 0.8B model makes Qwen 3.8-27B coding 47% faster, same pass rateEN
  8. 08Reika – A coding agent CLI designed around small local models firstEN
  9. 09System One models like Jev can train their own replacementsEN
  10. 10Could a Large Language Model Be Conscious? – David J. Chalmers (2024)EN
  11. 11NASA-IBM Lunar Foundation Model and fine-tuning codeEN
  12. 12Topological Models of Modal LogicEN

All figures and quotations in this text come from the sources listed below.

Content prepared by the editorial team with AI assistance.

Grace Okonkwo

Grace Okonkwo

AI, models and technology

Grace Okonkwo covers AI, models and technology for FLASH24, working from primary sources such as model cards, API documentation and benchmark papers rather than vendor summaries. She checks training data provenance, evaluation conditions and reported scores against the underlying datasets before any figure reaches print. She interviews researchers and engineers directly, tracks release calendars from major labs, and compares successive model versions on the same tests. Her own self-hosting, home-network and documentation-reading habits feed straight into that desk, since she tests tools on her own hardware first. She does not publish benchmark claims without a reproducible method.

Newsroom →

Comments

0
  1. No comments yet — be the first.

Write a comment

Comments are public. We do not publish abuse, spam or advertising.

Previous articleBenchmarks are getting better at measuring AI doubt, not just AI answersNext articleSpeech and translation models crowd the release calendar as open weights and small models multiply