Reka
★ 4/5 · Other
Reka builds multimodal frontier models that understand text, images, video and audio, offered via API and on-premise deployment for enterprise AI applications.
Pros
- Natively multimodal AI models supporting video, image, audio, and text processing
- Reka Edge (March 2026) provides fast inference at $0.10 per 1M input tokens
- Reka Flash 3 offers 128K context window for $0.35/million tokens (blended)
- flexible pay-as-you-go API pricing with no upfront costs
Cons
- Tier differentiation unclear
- Reka Edge and Flash 3 pricing suggests quality tradeoffs but feature parity not documented
- no free tier
- no public roadmap for additional models or features beyond existing Spark, Edge, Flash, Core tiers
Cost
Reka Edge: $0.10 per 1M input tokens; Reka Flash 3: $0.35/million tokens (blended); pricing range $0.10–$2.00 per 1M input tokens across model tiers (July 2026)
Verdict
Best for teams building multimodal AI applications who need cost-effective inference on a usage-based model, especially for video/image processing.
Visit Reka