rekaaiChat

Reka Flash 3

rekaai/reka-flash-3

66KContext Window

66KMax Output

Supported Protocols:reasoninginclude_reasoningtemperaturetop_ptop_kmax_tokensstopseedfrequency_penaltypresence_penalty

Online

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.

Capabilities

🧠 ReasoningText GenerationCode GenerationAnalysis & Reasoningmodels.reasoning

Technical Specs

Input Modality

Text

Output Modality

Text

Arch

—

Pricing

Pay per use, no monthly fees

Input Token< ¥0.001/1K Token

Output Token< ¥0.001/1K Token

Quick Start

from openai import OpenAI

client = OpenAI(
    base_url="https://api.uniontoken.ai/v1",
    api_key="YOUR_UNIONTOKEN_API_KEY",
)

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

print(response.choices[0].message.content)

FAQ

Reka Flash 3

rekaai/reka-flash-3

In< ¥0.001/1K

Out< ¥0.001/1K

Context Window66K

Max Output66K

Start Using →View Integration Docs

Related Models

Ready to get started?

Get 1M free tokens on registration, no monthly fees or minimum spend