whatmodel.sh / Flan-U-PaLM 540B

Flan-U-PaLM 540B

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

Flan-U-PaLM 540B is a closed-weights language model from Google, released October 2022. Its published size is 540B parameters. Licence: proprietary.

Catalogue identity notes

Google's Flan instruction-finetuned U-PaLM 540B (Chung et al., arXiv:2210.11416, Table 2/5). Never released. Comparison-only.

Model information sources

Release date ↗

51.5±1.4comparison only: no current run option verifiedfrom 4 benchmark results across 1 source

Score within the LMArena (text) comparison set only.

$ how it compares

The LLMs ranked either side of Flan-U-PaLM 540B, out of 877 with a score, then more from Google.

#modelscorewhereRAMreleased
406Qwen3 4B Alibaba52±1local~4GB2025-04
407Bielik 11B v2.0 Instruct SpeakLeash52±3local~8.5GB2024-08
408Flan-U-PaLM 540B Google52±1never released2022-10
409DeepSeek-Coder 33B Instruct DeepSeek51±1local~21GB2023-11
410Codestral 22B v0.1 Mistral AI51±1local~15GB2024-05
more from Google
11Gemini 4 Argon Google80±1api $4.00/M2026-09
26Gemini 3.8 Flash Google76±1api $1.50/M2026-09
34Gemini 3.7 Flash Google75±1api $1.50/M2026-08

see all LLMs · find the ones that fit your machine

$ what it's good at

math▁▂▃▆█▇▅▇▃▁▁▁ #185 of 523 * · 32% of its score · 1 test

Beats MiMo-V2.5 (310B-A15B), MiniMax-M2.1, Mistral Medium 3.5 (128B) here, though they rank higher overall. Trails OpenThinker3 7B here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
MGSM lab60.4% ✎▆·▂·▄▂▂▂▂█▄▆ beats 18/18medium · avg 47%▇▇▇▇▇▇▇▇▇▇▇▇▇ 32%sharp · at its level
other tests3 not in a task · 67% of its score
testresultvs the others in its rundifficultycountswhy
BIG-Bench Hard lab59.3% ✎▂▂▂▂·▅·▃▂█▃▃ beats 18/18medium · avg 60%▇▇▇▇▇▇▇▇▇▇▇▇ 30%sharp · at its level
MMLU lab74.1% ✎▁···▁··▃▁▅▅█ beats 18/18easy · avg 74%▇▇▇▇▇▇▇▇▇▇▇ 27%at its level
TyDiQA-GoldP (1-shot, EM, 8-language average) lab68.3% ✎▂··▂·▂▄▄·███ beats 18/18medium · avg 61%▇▇▇▇ 10%at its level

$ where to run it

No verified API or download recorded yet · add one

$ how the score was worked out · Scaling Instruction-Finetuned… 99%

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
Scaling Instruction-Finetuned Language Models — authors' own runs · lab-reported4CC BY 4.0
prior · pull toward the middle
weighted average= 51.5

$ 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