Flash 3

About Flash 3

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.

Model Specifications

💰 Real-World Cost Examples

Estimated monthly costs for common use cases

Rekaai Model Lineup

Compare all models from Rekaai to find the best fit

Similar Models from Other Providers

Cross-brand alternatives with similar capabilities

UI-TARS 7B

ByteDance

Input: $0.10

Output: $0.20

Context: 128k

Olmo 3 7B Instruct

Allen AI

Input: $0.10

Output: $0.20

Context: 66k

Mistral 7B Instruct v0.1

Mistral

Input: $0.11

Output: $0.19

Context: 3k

Ministral 3 8B 2512

Mistral

Input: $0.15

Output: $0.15

Context: 262k

💡 Cheaper Alternatives

Same Brand (Rekaai)

Cross Brand

🚀 Quick Start

Get started with Flash 3 API

from openai import OpenAI

client = OpenAI(
    base_url="https://api.provider.com/v1",
    api_key="YOUR_API_KEY"
)

response = client.chat.completions.create(
    model="rekaai/reka-flash-3",
    messages=[
        {"role": "user", "content": "Hello!"}
    ]
)
print(response.choices[0].message.content)

Continue Your Decision