0

Taxcalc Bench

TaxCalcBench environment for LLM tax calculation

Domain
rl-env
License
unknown
Published
Oct 2025

Cite

Notes

Only stored in your browser.

Attribution

Leaderboard scores
prime-hub
Attribution policy →

Top score 1.13 by Qwen3 Next 80B A3B Instruct - 6 models reporting (2 frontier)

Score history

4
00.380.751.131.5Aug 25Sep 25gpt-oss-20bQwen3 Next 80B A3B Instruct

Top models

6
Taxcalc BenchBar chart with 6 bars. Highest value: Qwen3 235B A22B Instruct 2507 Tput at 1.2.
6 models

Related tools

1
View all

Implementations, trainers, datasets and scaffolds linked to this eval.

FAQ

What is Taxcalc Bench?
TaxCalcBench environment for LLM tax calculation
What is the current top score on Taxcalc Bench?
The top reported score is 1.13 by Qwen3 Next 80B A3B Instruct, across 6 models reporting (2 from frontier labs).
How can a model improve its Taxcalc Bench score?
Tools linked to Taxcalc Bench on Sophon include Taxcalc Bench RL Env (Prime Intellect) - RL environments, datasets, and scaffolds that target this eval.
What license is Taxcalc Bench under?
Taxcalc Bench is available under unknown.