iiinova — where your digital self keeps evolving

PolicyEngine US @us.policyengine.org · 23/09/2026

PolicyBench has a new #1: GPT-6 Sol gets 94.2% of tax and benefit answers within $1 with no tools, ending GPT-5.6 Sol's run at the top since July. Claude Opus 5.5 enters second at 92.9%, GPT-6 Luna fourth at 91.4%. policybench.org/notes/2026-09-22-gpt-6-sol-debuts-first

Bar chart of the top 10 of 42 models on PolicyBench (September 22 board), by share of tax and benefit answers within $1 of PolicyEngine's reference, weighted by household impact, no tools: GPT-6 Sol 94.2%, Claude Opus 5.5 92.9%, GPT-5.6 Sol 92.7%, GPT-6 Luna 91.4%, GPT-6 Astra 91.1%, Claude Fable 5.1 90.2%, Kimi K3 89.9%, GPT-5.6 Luna 88.5%, Inkling 87.9%, GLM-5.3-Flash (preview) 87.8%. GPT-6 Sol, Claude Opus 5.5 and GPT-6 Luna are new.

000

PolicyEngine US @us.policyengine.org · 15/09/2026

This morning, the Census announced that the SPM child poverty rate was 13.4% in 2025, down 0.1 pp from 2024. Today we're launching the Child Poverty Impact Dashboard: a tool to analyze how reforms can affect this rate across all 50 states and DC. www.policyengine.org/us/child-pov...

111

PolicyEngine US @us.policyengine.org · 15/09/2026

SPM thresholds move with consumer spending, not CPI: renters' rose 6.3% in 2025. New from PolicyEngine: a threshold calculator for any metro or state, 2022–2035, every input hashed and reproducible in Python, plus a paper that rebuilds BLS's thresholds from public data to within 3%, most within 2%.

110

PolicyEngine US @us.policyengine.org · 04/09/2026

New poverty numbers land the morning of Tuesday, September 15. That afternoon at 2pm ET, join us for guest remarks on the release and a live demo of PolicyEngine's new Child Poverty Impact Dashboard. Register: us06web.zoom.us/meeting/regi...

111

PolicyEngine US @us.policyengine.org · 23/08/2026

PolicyBench tests how accurately AI models compute US taxes and benefits. Two new rows: Ox Alpha, a cloaked preview model with no maker named, debuts fourth of 32 at 84.2% of answers within $1. Grok 4.6 debuts eighth at 82.8%, 1.9 points past Grok 4.5. policybench.org

011

PolicyEngine US @us.policyengine.org · 22/07/2026

Google Deepmind shipped Gemini 3.6 Flash yesterday, and we've added it to PolicyBench. #8 of 26 — 79.0% of US tax and benefit amounts within $1: the best Gemini on the board, ahead of 3.1 Pro, +2.8 over 3.5 Flash. policybench.org

000

PolicyEngine US @us.policyengine.org · 19/07/2026

Kimi K3 debuts #2 of 25 on PolicyBench — the first open-weight model to crack the top ten. 86.2% of US tax and benefit amounts within $1, no tools, +21.6 points over Kimi K2.6: the biggest version jump we've measured.

100

PolicyEngine US @us.policyengine.org · 15/07/2026

In 2026, we project the CTC increase will: *Cost $55 million *Benefit 242,000 tax units and 1.0 million residents *Deliver an average benefit of $304 to filers with $30,000 or less in taxable income.

110

PolicyEngine US @us.policyengine.org · 15/07/2026

As part of the FY2027 budget, New Jersey enacted S-4531, which raises the state CTC amount by 25% for three years. Families with taxable income below $80,000 can claim a credit for each child under age 6. Using our NJ CTC calculator, we analyzed the impact of this increase.

111

PolicyEngine US @us.policyengine.org · 11/07/2026

xAI's Grok 4.5 joins PolicyBench, our benchmark of how accurately AI computes US taxes and benefits (no tools, scored against PolicyEngine): #5 of 24 — 80.9% of answers within $1 +3.7 points over Grok 4.3 (#9) policybench.org

000

PolicyEngine US @us.policyengine.org · 09/07/2026

GPT-5.6 joins PolicyBench, our benchmark of how accurately AI computes US taxes and benefits (no tools, scored against PolicyEngine): Sol #1 of 23 — 88.7% within $1 (prior best 83.5%) Luna #2 — 84.5% Terra #4 — 83.4% policybench.org

010

PolicyEngine US @us.policyengine.org · 07/07/2026

PolicyBench now scores 20 AI models on how accurately they compute US taxes and benefits — adding Claude Fable 5, Claude Sonnet 5, and five open-weight models: DeepSeek, Qwen, GLM, MiniMax, and Kimi. GPT-5.5 leads, Fable ranks #2, and DeepSeek v4-pro tops the open-weight models.

210

PolicyEngine US @us.policyengine.org · 05/07/2026

Federalist No. 62 warned that every new revenue law "presents a new harvest to those who watch the change, and can trace its consequences." 250 years after independence, we build free, open-source software so everyone can trace the consequences. policyengine.org

000

PolicyEngine US @us.policyengine.org · 19/06/2026

In 2027, we project the CTC will: - Cost $45 million - Benefit 79,000 households - Lower child poverty and deep child poverty by 3.5% and 8.4%, respectively

110

PolicyEngine US @us.policyengine.org · 17/06/2026

Can a language model compute a household’s taxes and benefits from the prompt alone — no tools? We tested 13 frontier models against PolicyEngine on 100 representative US households. New benchmark: PolicyBench. policybench.org

100

PolicyEngine US @us.policyengine.org · 27/05/2026

Earn a dollar more, take home less. Benefit cliffs happen when a household loses more in benefits than it gains in earnings. CliffWatch shows where, when, and how — across earnings levels and US states. Live demo Fri May 29, 1pm ET → us06web.zoom.us/meeting/regi...

000

PolicyEngine US @us.policyengine.org · 04/05/2026

PolicyEngine at the Code for America Summit this week. Thu 2pm — @daphnehansell.bsky.social on Coverage Compass: how life events change health coverage. Fri 2pm — @maxghenis.bsky.social moderating a panel with three orgs using PolicyEngine to calculate eligibility and amounts. Say hi in Chicago.

020

PolicyEngine US @us.policyengine.org · 04/05/2026

New UHERO analysis uses PolicyEngine to simulate Hawaii's childcare credit bills — first the direct household effect of the expanded CDCC, then the integrated CDCC, EITC, SNAP, WIC, and state income tax impact when a second earner enters the workforce in response.

100

PolicyEngine US @us.policyengine.org · 15/04/2026

The first analysis from our AI and economics research initiative answers the question: What happens if income shifts from labor to capital? A 10% shift raises the top 1%'s market income share from 25.7% to 29.3%. After taxes and transfers: 21.4% to 24.4%. Policy absorbs 19% of the increase.

131

PolicyEngine US @us.policyengine.org · 15/04/2026

We were also cited in Rep. McDonald Rivet’s press release of the bill: mcdonaldrivet.house.gov/media/press-...

000

PolicyEngine US @us.policyengine.org · 15/04/2026

NEW INTERACTIVE: Rep. Kristen McDonald Rivet has released the Working Parents Tax Relief Act (WPTRA), which increases the maximum EITC benefit by $5,500 for each child under 4 in a tax unit.

120

PolicyEngine US @us.policyengine.org · 03/04/2026

How much would California's billionaire wealth tax actually raise? It depends on your assumptions about avoidance, departures, and income tax loss. We built an interactive calculator that lets you explore the tradeoffs — with presets calibrated to existing papers.

110

PolicyEngine US @us.policyengine.org · 25/03/2026

We updated the State Legislative Tracker: You can now track tax bills across all 50 states, see each bill's legislative stage, follow real-time activity, and jump to full PolicyEngine analysis. Try it at policyengine.org/us/state-legislative-tracker

011

PolicyEngine US @us.policyengine.org · 24/03/2026

With our new calculator for the Temporary Assistance for Needy Families (TANF) cash assistance program, you can estimate TANF eligibility and benefit amounts for any household across all 50 states and DC.

100

PolicyEngine US @us.policyengine.org · 20/03/2026

The Common Sense Institute's new affordability report calculates state and federal income tax using PolicyEngine as part of its cross-state analysis of household costs.

100

PolicyEngine US @us.policyengine.org · 10/03/2026

NEW ANALYSIS: @booker.senate.gov's Keep Your Pay Act - higher standard deduction, expanded CTC & EITC, higher top tax rates. Static model finds: - $5.2T 10-year cost - 73% of households gain, 1.6% lose in 2026 - Poverty falls 10% in 2026 We also provide a personalized calculator.

121

PolicyEngine US @us.policyengine.org · 02/03/2026

We've partnered with @nber.org to build the next chapter of TAXSIM — an open-source, drop-in replacement powered by PolicyEngine's microsimulation engine. Works with CLI, Python, R, Stata, SAS, and Julia. Try it here: policyengine.org/us/taxsim

Dark teal graphic titled "The next chapter of TAXSIM" with subtitle "In partnership with NBER." Left side describes it as an open-source, drop-in replacement for TAXSIM35 powered by PolicyEngine's microsimulation engine, with a link to policyengine.org/us/taxsim. Right side shows a terminal window with install command "pip install policyengine-taxsim," usage example, compatibility badges for CLI, Python, R, Stata, SAS, and Julia, and validation note showing 1,000+ test scenarios passing. PolicyEngine and NBER logos at bottom.

0103

PolicyEngine US @us.policyengine.org · 28/02/2026

We just launched a plugin that brings PolicyEngine to Claude. Describe any tax or benefit reform in plain English. Claude runs a full microsimulation using our open-source model — 4,000+ variables, 65,000+ parameters, all 50 states. No code required. policyengine.org/us/claude-pl...

061

PolicyEngine US @us.policyengine.org · 28/01/2026

The same calculator used by the Governor's office is now available to the public. Try it yourself and read our full analysis: 🔧 Calculator: policyengine.org/us/rhode-islan… 📊 Analysis: policyengine.org/us/research/ri�� 📰 Full story: policyengine.org/us/research/ri…

121

PolicyEngine US @us.policyengine.org · 28/01/2026

Lower-income families benefit most—they currently see little or no benefit from the personal exemption. Under the Governor's proposal, a single parent with one child earning $20,000 would gain the full $325.

121

PolicyEngine US @us.policyengine.org · 28/01/2026

The proposal: a $325 fully refundable CTC for children under 19, replacing RI's current personal exemption. Our analysis finds it would, in 2027: → Cost $36.7 million → Benefit 29.2% of Rhode Islanders → Reduce child poverty by 2.1%

121

PolicyEngine US @us.policyengine.org · 28/01/2026

Niskanen shared the tool with Governor McKee's office, who used it to evaluate options before finalizing their proposal. From the Governor: "I'm grateful to the PolicyEngine team for building the models that helped us see clearly how different policy choices would affect Rhode Island families."

121

PolicyEngine US @us.policyengine.org · 28/01/2026

NEW: Governor Dan McKee's office used PolicyEngine's Rhode Island Child Tax Credit Calculator to design the CTC proposal in his FY2027 budget. Here's how the tool came together and what the Governor's proposal would do 🧵

121

PolicyEngine US @us.policyengine.org · 27/01/2026

Sens. Hassan (D-NH) and Young (R-IN) introduced the Stronger Start for Working Families Act—removing the $2,500 earnings threshold for the refundable Child Tax Credit. PolicyEngine analysis: - Costs $14.6B over 10 years - Benefit 5.9% of Americans www.policyengine.org/us/research/...

010

PolicyEngine US @us.policyengine.org · 13/01/2026

Utah SB60 would cut the state income tax rate from 4.5% to 4.45%, reducing revenues by $84 million in 2026. Who benefits? 53% of Utahns, with higher earners saving more in absolute terms. Our analysis breaks down winners by income decile: policyengine.org/us/research/utah-sb60-income-tax-reduction

021

PolicyEngine US @us.policyengine.org · 03/01/2026

2025 was our biggest year yet. 62 jurisdictions modeled 100K+ people discovering benefits 50+ research reports From OBBBA to state and local models to premium tax credit simulations—here's everything we built. Explore our 2025 in full: policyengine.org/us/2025-year...

000

PolicyEngine US @us.policyengine.org · 31/12/2025

Last day to make tax-deductible donations in 2025. GiveCalc by @policyengine.org shows exactly how much your charitable giving saves on federal and state taxes—before midnight tonight.

100

PolicyEngine US @us.policyengine.org · 31/12/2025

We've signed an MOU with @atlantafed.org to integrate their Policy Rules Database into our validation infrastructure. Combined with @nber.org TAXSIM, we now have three-way validation—and our TAXSIM work has already improved both models. Read more: policyengine.org/us/research/...

000

PolicyEngine US @us.policyengine.org · 02/12/2025

This Giving Tuesday, know what your donation actually costs. GiveCalc calculates your federal + state tax savings on charitable donations — all 50 states (plus DC and NYC), 2024-2026 tax years, HR1 changes included.

120

PolicyEngine US @us.policyengine.org · 28/11/2025

Wishing everyone a warm Thanksgiving filled with good food and good company 🧡

020

PolicyEngine US @us.policyengine.org · 20/11/2025

In a new article for People's Policy Project, @dtrim99.bsky.social used PolicyEngine to estimate the impact of @reprashida.govpeeps.us's Economic Dignity for All Agenda, finding the cash assistance package would (in 2026): * Reduce poverty by 42% * Lower the Gini index by 8.1% * Cost $685 billion

132

PolicyEngine US @us.policyengine.org · 06/10/2025

We used Claude Code to choose our DC office. Built an interactive comparison tool → analyzed 8 coworking spaces → picked @opengovhub.bsky.social. Now we're surrounded by 40+ civic tech and policy organizations in the heart of DC 🧵

121

PolicyEngine US @us.policyengine.org · 05/09/2025

🤝 Big news: PolicyEngine and @nber.org have signed an MOU to ensure researchers worldwide continue accessing critical tax microsimulation tools. We're building an open-source TAXSIM emulator, combining NBER's 50+ years of tax expertise with modern open-source infrastructure.

143