ZMIME

Dhīr 5B MoE vs Olala Mini 7A1B vs EKA-1

Compare documented capabilities and access. A dash means a comparable figure is unavailable.

Model specifications
SpecificationDhīr 5B MoEZenteiQOlala Mini 7A1BOVHcloudEKA-1Soket AI
TaskMultilingual assistantMultilingual language generationMultilingual reasoning and coding
Published capabilitiesMixture-of-experts assistant supporting English and ten Indian languages for offline deployment.Compact mixture-of-experts model combining hybrid attention and Mamba layers.Trained on about 7.5 trillion tokens across English and all 22 scheduled Indian languages.
Model size5.09B total / 1.79B active7B total / 1B active26.3B parameters
AccessOpen weights for on-premises or offline useDownloadable checkpoint for local inference or fine-tuningReleased through the IndiaAI Mission
SourcesPublisher documentation · Additional documentationModel weights · Additional documentationPublisher documentation
InputsTextTextText
Context window16,384 tokens32,000 tokens—
Input price per million tokens———
Output price per million tokens———
WeightsOpen weightsOpen weights · Apache-2.0Weights unconfirmed
ReleasedNov 3, 2026Nov 19, 2026Nov 3, 2026