Skip to main content
Models/Reka/Reka Flash

Reka Flash

Reka 🧠 Language Model

Reka Flash, fast and cost-efficient multimodal model for everyday tasks. Accepts text, image, video, and audio. 128K context.

⏹ Retired on 2026-10-02

Specifications

Input modalitiestext, image, video, audio
Output modalitiestext
Relative speed●●●●○ Fast
Deprecated on2026-10-02
Sunset on2026-10-02
Context window128,000 tokens
Max output8,192 tokens
Vision inputYes
Tool callingNo
Streaming outputYes
Input price$1.40 / 1M
Output price$3.50 / 1M

Pricing

$1.40 / 1M input tokens  ·  $3.50 / 1M output tokens

Platform credit pricing — live rates on the Models page.

More from Reka

R
Reka Core
Reka's most capable multimodal model, accepts text, images, video, and audio input. 128K context, strong reasoning and analysis.
🧠 Language Model
R
Reka Flash 3.1
Reka's latest fast model, excellent balance of speed and capability with multimodal input (text, image, video, audio). 128K context.
🧠 Language Model
R
Reka Edge
Reka's smallest and cheapest model, ideal for simple tasks and high-volume use. Multimodal input, 128K context.
🧠 Language Model
R
Reka Edge 2
7B SoTA vision language model optimized for efficiency.
🧠 Language Model