01 / SUBMIT
Bring the product
A team submits a working AI product, not just a deck, ticker or roadmap.
Working demo required
Product-first AI launchpad on Solana
Proof before price.
BENCH puts working AI products in front of real users before launch. Teams earn access by proving people can use the product, complete real tasks and create measurable value.
Test · verify · score · unlock
01 / Mechanism
A launch starts with a working product.
BENCH replaces instant launch with a public validation phase. The product creates evidence first; the market comes later.
02 / TEST
Open the bench
Real users complete product tasks, find bugs and pressure-test the experience.
72H public phase
03 / SCORE
Measure usage
Task success, reliability, feedback and repeat usage become a public score.
Visible results
04 / UNLOCK
Earn access
Products that clear the proof threshold move forward. Weak ones return to building.
Proof > hype
02 / Product first
The market has enough promises.
Products before promises.
BENCH gives working products a place to prove themselves before attention turns into speculation.
A usable product comes before token access.
Testing creates an inspectable record of what works.
Teams iterate from real feedback before launch pressure begins.
03 / Public testing
72 hours to prove it.
The test window is short, visible and centered on real product use — not follower counts.
BENCH TEST / B-0427PUBLIC TEST ACTIVE
AI research agent — complete real tasks with real users.
71:59:28
Verified sessions428
Task success86%
Repeat usage41%
Session #428712s
Summarize a research paper✓ CompletedSession #428828s
Analyze token design✓ CompletedSession #428941s
Find market opportunities✓ Completed
Testing is part of the launch.
Build, test, verify, improve — then move forward. Product validation stays visible instead of hiding behind a roadmap.
Public tasks reveal strengths and failure points.
The test history stays attached to the project.
Iteration happens before market pressure.
04 / BENCH Score
A score built from usage.
Follower count does not tell you whether an AI product works. BENCH Score is built from verified testing signals.
BENCH SCORE / B-0427
82
PASSED / 70+ REQUIRED
Task success88
Quality76
Reliability81
Repeat use74
05 / Contribution
Your contribution counts.
BENCH rewards useful participation instead of nothing but speed. Users help products improve while building a visible contribution history.
Complete tasks
Run actual workflows inside the product.
Report bugs
Surface reproducible issues that matter.
Give feedback
Submit useful structured product insight.
Use again
Repeat usage is stronger evidence than one click.
06 / Signal
Usage beats hype.
Attention can bring people to a product. Only usage shows whether they stay, complete the task and return.
Completed flows beat impressions.
Repeat users beat follower screenshots.
Verified feedback beats vague social sentiment.
Don't launch the promise. Prove the product.
BENCH gives AI products a place to earn attention through real usage before the market gets involved.
07 / FAQ
Products before promises.
What is BENCH?
BENCH is a product-first AI launchpad on Solana. Working products enter public testing, generate verifiable usage data, receive a BENCH Score and can earn launch access.
What happens during a public test?
Users complete real product tasks, report bugs, submit structured feedback and generate measurable product-use signals.
What is BENCH Score?
A product score built from testing signals such as successful task completion, output quality, reliability and repeat usage.
Does every product get launch access?
No. Projects need to clear the requirements of their test phase. Products that miss the threshold return to iteration.