whatmodel.sh / WizardLM-2 8x22B

WizardLM-2 8x22B

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

WizardLM-2 8x22B is an open-weights language model from Microsoft, released April 2024. Its published size is 141B parameters (39B active per token). Licence: apache-2.0.

Catalogue identity notes

Comparison model from the Epoch AI internal-evaluation CSVs (Epoch release date 2024-04-15). WizardLM-2 8x22B (April 2024), a fine-tune of Mixtral 8x22B. 141B parameters and Apache-2.0 per the ollama model page/licence layer; 39B active is the Mixtral 8x22B figure (Mistral: '39B active parameters out of 141B'). Microsoft's Hugging Face repo is no longer publicly accessible (HTTP 401, observed 2026-09-28).

49.5±1.3#351 of 719 LLMsfrom 7 benchmark results across 6 sources

$ how it compares

The LLMs ranked either side of WizardLM-2 8x22B, out of 828 with a score, then more from Microsoft.

#modelscorewhereRAMreleased
420GLM-4 (0116) Z.ai50±1comparison only2024-01
421Llama 3 70B Instruct Meta50±0local~24GB2024-04
422WizardLM-2 8x22B Microsoft50±1local~91GB2024-04
423Nova Pro Amazon50±1api $1.40/M2024-12
424WizardCoder 33B V1.1 WizardLM50±1local~22GB2024-01
more from Microsoft
252Phi-4-reasoning-plus 14B Microsoft59±1local~10GB2025-04
261Phi-4-reasoning 14B Microsoft58±1api $0.22/M +local~10GB2025-04
365Phi-4 14B Microsoft53±1api $0.087/M +local~10GB2024-12

see all LLMs · find the ones that fit your machine

$ what it's good at

coding▁▁▁▃▆██▇▆▄▃▁ #307 of 458 * · 28% of its score · 1 test

Beats Solar Open 100B (102B-A12B), GLM-4.7-Flash, Qwen3 4B Instruct 2507 here, though they rank higher overall. Trails DeepSeek LLM 67B Chat, WizardLM 70B V1.0, Yi 1.5 9B Chat here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
Aider code editing (133 Exercism Python exercises, two attempts)44.4% ✎▁·▁▂▄▃▆▄█▆▂▂ beats 12/62medium · avg 54%▇▇▇▇▇▇▇▇▇▇▇ 28%at its level
math▁▂▃▅▆▅▄█▄▁▁▁ #301 of 398 * · 31% of its score · 1 test

Beats GPT-4 (0613), Athene-70B, Llama 3 70B Instruct here, though they rank higher overall. Trails Gemma 7B (base), DeepSeek R1 Distill Qwen 1.5B, CodeGemma 7B (v1.0) here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
MATH level 525.7% ✎▂▂▅▆▄█▆▃▅▂▃▄ beats 25/92hard · avg 32%▇▇▇▇▇▇▇▇▇▇▇▇ 31%sharp · at its level
other tests5 not in a task · 41% of its score
testresultvs the others in its rundifficultycountswhy
GPQA Diamond43.4% ✎▁▁▁▁▂▄▆▄▅▇█▅ beats 47/208hard · avg 40%▇▇▇▇▇▇▇ 16%sharp
CPTUB average (SpeakLeash, 0-5 judge scale)74.0% ✎▁▂▂▂▃█▃▆▅▄▂▃ beats 65/95easy · avg 66%▇▇▇▇ 10%at its level
Polish EQ-Bench v2 (SpeakLeash)69.6% ✎▁··▁▁▂▂▆▅▆█▆ beats 51/70medium · avg 64%▇▇▇ 8.0%at its level
Judgemark v2 (LLM-as-judge for creative writing)55.4% ✎▃▃··▃▅▅···█▃ beats 2/10medium · avg 55%▇▇ 5.6%at its level
Vectara hallucination leaderboard, 2025 dataset (HHEM-2.3, factual consistency rate)88.3% ✎▁▁▂▃▄▄█▆▃▂▂▁ beats 17/140very easy · avg 94%▇ 0.9%blunt · maxed out · noisy

$ where to run it

Ollama · recommended

version

  1. Install Ollama

    download it (macOS / Windows) ↗

    Linux
    curl -fsSL https://ollama.com/install.sh | sh
  2. Run WizardLM-2 8x22B · needs ~85.1 GB · from ollama.com
    chat in the terminal
    ollama run wizardlm2:8x22b
$ how the score was worked out · Epoch AI — internal evaluatio… 47%, Aider code editing benchmark… 28%, SpeakLeash CPTUB (Complex Pol… 10% and 3 more

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
Epoch AI — internal evaluations2▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 47%CC BY 4.0
Aider code editing benchmark runs (old edit leaderboard)1▇▇▇▇▇▇▇▇ 28%Apache-2.0
SpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)1▇▇▇ 10%MIT
SpeakLeash Polish EQ-Bench leaderboard1▇▇ 8%Apache-2.0
Judgemark v2 results (EQ-Bench)1▇▇ 6%MIT
Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository)1▇ 1%Apache-2.0
prior · pull toward the middle▇ 1%
weighted average= 49.5

$ 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