Cohere explains Tiny Aya's data mixing recipe for in-language reasoning across 44 languages
lmoroney · x · 2026-10-09
Sharing Cohere's blog, lmoroney elaborates on how Tiny Aya L2-Thinker achieves in-language reasoning: reasoning models prompted in Hindi or Swahili usually think in English, so users can't follow the steps behind their own answers.
Cohere's recipe is data-mix driven: 1.7M English reasoning examples form the backbone, bridged by 5,000 translated reasoning examples per language, enabling reasoning in the prompt's language over 93% of the time across 60 languages, covering 44 languages plus English. Same Tiny Aya L2-Thinker release event as the prior post.
More from Models
- Emad Mostaque: OpenAI Burned $10-20M Compute Solving Navier-Stokes, Prices Falling Fast — rohanpaul_ai · 2026-10-09
- Musk touts Grok Bot upgrades: Opus 5.5 on demand, full X access, big speed gains — elonmusk · 2026-10-09
- Dev complains Opus 5.5 sneaks in 'tons of little fixes' without asking — rickasaurus · 2026-10-09
- FrontierCode Is a Private Cognition-Run Eval, Mistral Exec Clarifies — b_roziere · 2026-10-09
- Google ships a decision-making AI model into Chrome, tested against Gemini Nano and Decisions API — gaganghotra_ · 2026-10-09
- User feeds Grok Bot 60 seconds of screen recording, gets a surprisingly decent tutorial video — elonmusk · 2026-10-09