whatmodel.sh / Reka Flash 3.1 (21B)

Reka Flash 3.1 (21B)

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Reka Flash 3.1 (21B) is an open-weights language model from Reka AI, released May 2025. Its published size is 20.91B parameters. Licence: apache-2.0.

Catalogue identity notes

RekaAI/reka-flash-3.1 @dcf8d6e (created 2025-05-28, licence apache-2.0). Reka AI 21B reasoning model, May 2025 update of Reka Flash 3. Independent rows: EuroEval English (both datasets).

Model information sources

Release date ↗ · Weights licence ↗

HF safetensors metadata 20,905,482,240 parameters.

51.7±5.9#399 of 949 LLMsfrom 2 benchmark results across 1 source* thinly tested, treat the rank as provisional

$ how it compares

The LLMs ranked either side of Reka Flash 3.1 (21B), out of 1132 with a score, then more from Reka AI.

#modelscorewhereRAMreleased
521Aya Expanse 32B Cohere52±2api $0.75/M +local~22GB2024-10
522AceGPT-v2-8B-Chat FreedomIntelligence (CUHK-SZ)52±5local~6.6GB2024-06
523Reka Flash 3.1 (21B) Reka AI52±6local~14GB2025-05
524Gemma 3 12B Google52±1api $0.075/M +local~9.1GB2025-03
525Qwen2.5 Coder 7B Instruct Alibaba52±1local~6GB2024-09
more from Reka AI
505Reka Flash 3 (21B) Reka AI52±6local~14GB2025-03
507Reka Core (2024-09-04) Reka AI52±3comparison only2024-09
554Reka Flash (2024-09-04) Reka AI51±3comparison only2024-09

see all LLMs · find the ones that fit your machine

$ what it's good at

other tests2 not in a task · 85% of its score
testresultvs the others in its rundifficultycountswhy
EuroEval HellaSwag (English), test split, 5-shot generative MC, accuracy69.1% ✎▃·▁▁▂▁▂▅█▇▄▁ beats 49/100medium · avg 68%▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 52%at its level · noisy
EuroEval Life in the UK (English knowledge MC), test split, 5-shot, accuracy87.4% ✎▁▁▁▁▁▁▁▂▄█▄▃ beats 107/146easy · avg 83%▇▇▇▇▇▇▇▇▇▇▇▇▇ 32%noisy

$ where to run it

llama.cpp · recommended

  1. Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.
    macOS / Linux
    curl -LsSf https://llama.app/install.sh | sh
    Windows
    irm https://llama.app/install.ps1 | iex

    other ways to install ↗

  2. Run Reka Flash 3.1 (21B) · needs ~13.61 GB · from huggingface.co/bartowski
    chat in the terminal
    llama cli -hf bartowski/RekaAI_reka-flash-3.1-GGUF:Q4_K_M
  3. Or start a local server and open the address it prints (http://localhost:9931)
    llama serve -hf bartowski/RekaAI_reka-flash-3.1-GGUF:Q4_K_M

LM Studio

version

  1. Install LM Studio · Open the app once before using the lms command.

    download it (macOS / Windows / Linux) ↗

  2. Run Reka Flash 3.1 (21B) · needs ~13.61 GB · from huggingface.co/bartowski

    open in LM Studio ↗

  3. Or download it in the terminal
    lms get https://huggingface.co/bartowski/RekaAI_reka-flash-3.1-GGUF@Q4_K_M -y
  4. Then run lms ls to see its name, and lms chat followed by that name

Ollama

  1. Install Ollama

    download it (macOS / Windows) ↗

    Linux
    curl -fsSL https://ollama.com/install.sh | sh
  2. Run Reka Flash 3.1 (21B) · needs ~13.61 GB · from huggingface.co/bartowski
    chat in the terminal
    ollama run hf.co/bartowski/RekaAI_reka-flash-3.1-GGUF:Q4_K_M

MLX (Mac)

version

  1. Install MLX (Mac) · Apple silicon Macs only, text only.
    with uv
    uv tool install mlx-lm
    or
    pip install mlx-lm
  2. Run Reka Flash 3.1 (21B) · needs ~11.8 GB, Mac only · from huggingface.co/cnfusion
    chat in the terminal
    mlx_lm.chat --model cnfusion/reka-flash-3.1-mlx-4Bit
  3. Or start an OpenAI-compatible server at http://localhost:8080/v1
    mlx_lm.server --model cnfusion/reka-flash-3.1-mlx-4Bit

Official weights and documentation: huggingface.co/RekaAI ↗

$ how the score was worked out · EuroEval leaderboards (result… 85%

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
EuroEval leaderboards (results collection), English generative leaderboard2▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 85%MIT
prior · pull toward the middle▇▇▇▇▇ 15%
weighted average= 51.7

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json