Model

Reka Core

CLOSED

Reka AI · Reka family · released Apr 15, 2024

Reka's frontier multimodal model (text/image/video/audio); approached GPT-4V/Claude-3-Opus class on multimodal + language benchmarks at launch.

ReasoningCodingVisionFunction callingTool useAgentic

909.9

Elo · rank #233

Parameters

Undisclosed

Active params

—

Context

128K tokens

Architecture

Dense transformer, PaLM-like (no parallel layers); full architecture undisclosed

License

Proprietary

Languages

32+

API price (in/out)

$2 / $6

Modalities

text · vision · audio · video

Benchmark results

Bar shows position within the tracked field; marker = field best

GPQA DiamondReasoning38.2%#228

best: GPT-5.6 · 94.6%

GSM8KMath92.2%#30

best: Llama 3.1 405B · 96.8%

HumanEvalCoding76.8%#62

best: Claude Opus 4.5 · 99.4%

MMLUKnowledge83.2%#72

best: OpenAI o3 · 92.9%

MMMUVision56.3%#81

best: Claude Fable 5 · 89.3%

PubMedQAKnowledge74.6%#6

best: Palmyra Med 70B · 79.6%

VQAv2Vision78.1%#13

best: Reka Edge (2026) · 88.4%

XWinogradReasoning86.8%#2

best: Sarashina2-70B · 91.8%

Run it locally

Closed weights — available via API only. No local deployment.

Input / M tok

Output / M tok

Reka family

Elo progression across releases

Reka Edge (7B)Feb 2024378.0 Reka Flash (21B)Feb 2024633.5 Reka CoreApr 2024909.9 Reka Flash 3Mar 20251241.4 Reka Edge (2026)Mar 20261698.3

Resources

DOCSdocs.reka.ai APIAvailable via Reka AI API

API price $2/$6 · each benchmark row carries its own source badge (see methodology)