Opper

Opper published an updated explainer on how LLM leaderboards work in 2026, breaking down three primary methodologies: the Artificial Analysis Intelligence Index (version 4.1.1, a composite of nine standardized evaluations run by AA with pass@1 scoring), Arena (formerly LMArena) human-preference pairwise battles convert The article also introduces Opper's own LLM leaderboard, which layers per-route prices, measured speed, and hosting information on top of the Artificial Analysis indices. The updated date is August 17, 2026. This is Opper-authored methodology content rather than a model or API release, but it is relevant provider-side

Novita AI published a partnership announcement stating that its inference is now available through Opper AI's EU-hosted gateway, giving European agent builders access to more than 80 open-weight models via a single API. The collaboration combines Novita's inference infrastructure with Opper's European hosting and routi For Opper as a provider, this is relevant ecosystem news because it documents a concrete third-party integration with Opper's EU gateway and a specific catalog claim (80+ open-weight models). The supplied excerpt confirms the partnership framing, the EU hosting positioning, and the single-API access model, all of which