we made distributed inference verifiable with <1% overhead.
verification is critical for any distributed system. in a trustless network, actors may swap your 70B model for a cheaper 8B one to cut costs.
until now, maintaining inference integrity meant either doubling your cost (redundancy) or exploding your latency (zkp).
we created veri: an on-chain verification layer light enough for high-throughput frameworks like Parallax. it hits the economic sweet spot through architectural elegance:
1. commit-sample-verify
we don't prove every step; we check a random slice using game theory. workers commit to their work before the audit. cheating becomes statistically irrational, allowing a 1% sample to secure the entire sequence.
2. simultaneous execution
inference and verification happen simultaneously on the same worker pool. we don't need a separate "verifier set", so compute utilization stays high.
find out more about the architecture and benchmarks:
paper: https://t.co/0nEhVzNQUe
blog: https://t.co/ULpLUqtGmz
Running large language models locally used to be reserved for the most technical users.
With Parallax, we redefine where AI runs.
Build your own AI lab today, and compete for a DGX Spark + 7 Mac Minis.
The challenge starts now.
We're thrilled to announce our strategic partnership with @Zai_org.
Experience GLM-4.6 on Parallax now with unparalleled coding prowess, an expansive 200K token context for complex tasks, and superior agentic reliability.
We'll work closely to bring more possibilities and efficiency to your own AI lab.
Towards open intelligence.
Thrilled to partner with @Gradient_HQ in bringing open intelligence to everyone. GLM-4.6 now runs on Gradient’s Parallax, unlocking its power for more users worldwide. Looking forward to what we build together next!
Open-source models are no longer overshadowed by closed ones.
Congrats to our friends at @Kimi_Moonshot for moving us toward Open Intelligence.
Experience Kimi K2 Thinking with Parallax now.
Gradient is built for this shift.
Parallax is your ideal local AI engine for building x402 agents.
Win a $10,000 pool co-sponsored with our friends at @trendsdotfun.
The trends are good.
https://t.co/A6PmuPraLI
Gradient invites the world to assemble, scale, and evolve intelligence on consumer devices.
We are researching into:
-P2P Communication & Orchestration
-Distributed ML
-Open platform
An open stack powered by everyday hardware, free from centralized clouds.
The stack for collective intelligence is taking shape piece by piece.
Open AGI is the destination, and we’re building toward it every day.
Learn more about Gradient.
https://t.co/hbg97yHXSP
Good day from Open Intelligence Day Singapore!🇸🇬
We’re all set for a full day of keynotes, panels, and demos shaping open AGI.
Follow this thread for real-time updates throughout the day. 👇
Blockchain is much bigger than just DeFi and crypto.
Yesterday, @AgriDexPlatform and @dawninternet showed how they power agriculture and digital infrastructure on Solana.
We’re proud to give our founders a voice in policy, it’s vital for our industry’s future.