Flash 3 Check detailed information and pricing for AI models

Context Length 32,768 tokens, rekaai from provided

32,768
Context Tokens
$0.01
Prompt Price
$0.01
Output Price
10/16
Feature Support

Model Overview

Reka Flash 3 is a general-purpose, instruction-tuned large language model with 21 billion parameters, developed by Reka. It excels at general chat, coding tasks, instruction-following, and function calling. Featuring a 32K context length and optimized through reinforcement learning (RLOO), it provides competitive performance comparable to proprietary models within a smaller parameter footprint. Ideal for low-latency, local, or on-device deployments, Reka Flash 3 is compact, supports efficient quantization (down to 11GB at 4-bit precision), and employs explicit reasoning tags ("<reasoning>") to indicate its internal thought process. Reka Flash 3 is primarily an English model with limited multilingual understanding capabilities. The model weights are released under the Apache 2.0 license.

Basic Information

Developer
rekaai
Model Series
Other
Release Date
2025-03-12
Context Length
32,768 tokens
Variant
standard

Pricing Information

Prompt Tokens
$0.01 / 1M tokens
Completion Tokens
$0.01 / 1M tokens

Data Policy

Terms of Service

학습 정책

1

Supported Features

Supported (10)

Top K
Seed
Frequency Penalty
Presence Penalty
Repetition Penalty
Min P
Logit Bias
Logprobs
Top Logprobs
Reasoning

Unsupported (6)

Image Input
Response Format
Tool Usage
Structured Outputs
Web Search Options
Top A

Other Variants

Actual Usage Statistics

#349
Out of 353 total models
51.15K
Total Tokens Last 30 Days
51.15K
Daily Average Usage
-
Weekly Usage Change

Usage Trend for the Last 30 Days

Similar Price Range Models

LFM 7B
liquid
32,768 tokens
$0.01 / $0.01
Llama 3.2 1B Instruct
meta-llama
131,072 tokens
$0.01 / $0.01