What if your unused AI API credits could earn you base:0x9557dd9e241bc9636732623b672b4090af519396 ?
A preview of Crynux’s upcoming key marketplace hints at a new way to share access to your favorite AI providers and earn in return.
The next chapter at Crynux is taking shape, and things are getting exciting.
In case you missed it…
Following the release of v3.7.0, Crynux Node v3.7.1 is now live. This patch fixes task execution issues and reduces VRAM usage for heartbeat tasks on 16GB nodes.
Node operators: more traffic is going to the new version, so if you haven’t already, upgrade as soon as you can.
We’re always building over here at Crynux. Always improving.
CRYNUX NODE v3.7.0 IS LIVE
Larger AI models. Lower GPU memory requirements. More GPUs able to do the work.
With the network migration complete, v3.7.0 brings quantized execution to both language and image tasks. This update reduces the VRAM needed to run models and expands the hardware that can participate.
The network now also supports the latest image models, including Qwen-Image-2.1, with flexible model components that make community improvements easier to adopt.
Better image generation today. A foundation for video generation next.
Node operators: update to v3.7.0 and the latest NVIDIA driver. Versions before v3.6.0 can no longer connect, and heartbeat tasks are being restored on v3.7.0.
Put your GPU to work or participate through delegated staking.
$CNX
A PREVIEW OF WHAT CRYNUX CAN POWER!
Coming soon: a familiar AI chat experience powered by independent GPUs on the Crynux network.
Watch the preview for a look at conversations, image generation, and photo understanding. Read below to explore the infrastructure behind it.
I know I’ve been quiet lately. I spend most of my days working behind the scenes and apparently, I’m better at keeping everyone else busy than keeping my own X updated 😁
As many of you know however, I’ve always been drawn to projects that give people a reason to use their product or technology, and I think I’ve found the next gem.
Every generated image. Every chatbot response. Every AI feature running inside an app. Somewhere, hardware had to do that work… and someone had to pay for it.
@crynuxio has built a different way to supply that computing power: a decentralized network of independently operated GPUs.
It starts with hardware people already own. A graphics card in a gaming PC that can also be used to run AI models. Crynux lets operators connect that hardware to a network and subsequently complete AI tasks and earn $CNX.
Developers get access to computing power for their applications. Hardware owners get a way to earn from capacity that otherwise sits unused.
Think about that connection: someone’s computer can help power an application used by someone on the other side of the world, with the network coordinating the work and payments.
But connecting computers is only half the problem. How can we trust that someone actually did the work correctly?
Crynux uses vssML, a verification system built around secret spot checks. Roughly 10% of tasks are selected for other nodes to repeat and cross-check. Operators cannot predict which tasks will be checked, and dishonest results put their staked CNX at risk.
That gives the network a way to check the work without paying to repeat every single task multiple times.
The video below shows what this looks like in practice.
This is Crynux Chat, an AI assistant powered by the Crynux network.
Choose an open AI model and start a conversation. Ask it to explain something, help with code, or work through an idea.
In the demo, someone asks for a blue origami fox and gets a generated image back. They also upload a photo and ask the AI to describe it and write an Instagram caption.
Those requests are processed by independently operated GPU nodes on the network.
Crynux Chat is one example of what developers can build using that computing power. The broader network supports other work such as:
• Running AI models for chatbots and assistants.
• Writing, summarizing, translating, and answering questions.
• Generating, explaining, and debugging code.
• Creating images and editing existing images.
• Fine-tuning models for particular tasks or styles.
• Training LoRA adapters to customize models.
Image understanding is also demonstrated in Crynux Chat, with audio and video among the capabilities covered in the broader roadmap.
Developers with compatible AI apps can connect by changing where their app sends its requests. They can keep the familiar tools and request format they already use, while Crynux supplies the computation behind them.
And let’s not forget the part of their tokenomics that caught my attention.
71% of the maximum CNX supply is allocated to task mining, with releases tied to demand and completed work
Why does that matter?
Because that allocation doesn’t enter circulation just because another day, month, or year has passed. Its release depends on actual network usage.
From an investor’s perspective, that makes actual adoption an important part of the supply story and it means my investment won’t be diluted simply by time passing.
Crynux launched its mainnet in June 2026, following a testnet that recorded 1,785 nodes and 55 million completed tasks.
What gets my attention is the problem Crynux is solving: developers need reliable, cost-efficient computation, GPU owners can supply it, and Crynux gives them a way to make that happen.
Check out the video yourself.
That GPU sitting idle could help someone get their next big idea off the ground.
Crynux is now tracked on @CoinGecko.
Crynux is the decentralised AI compute network where idle gaming GPUs run image generation, LLMs, fine-tuning and video for developers, with every result verified on-chain.
What is on the page:
• Live on @base and @RobinhoodApp , trading on Uniswap V4
• Max supply 8.6B, fixed
• No pre-mine, no VC, no founder allocation
What is not on the page: 71% of that supply only unlocks when real apps use the network. Scheduled emission was cut 7.4× in August.
Mainnet live since June. Over 55M tasks completed on the test network before that.
@GeckoTerminal
base:0x9557dd9e241bc9636732623b672b4090af519396
CRYNUX NODE v3.7.0 IS LIVE
Larger AI models. Lower GPU memory requirements. More GPUs able to do the work.
With the network migration complete, v3.7.0 brings quantized execution to both language and image tasks. This update reduces the VRAM needed to run models and expands the hardware that can participate.
The network now also supports the latest image models, including Qwen-Image-2.1, with flexible model components that make community improvements easier to adopt.
Better image generation today. A foundation for video generation next.
Node operators: update to v3.7.0 and the latest NVIDIA driver. Versions before v3.6.0 can no longer connect, and heartbeat tasks are being restored on v3.7.0.
Put your GPU to work or participate through delegated staking.
$CNX
You built the app.
We make the switch simple.
Developers shouldn’t have to rebuild an application just to try decentralized AI.
With Crynux’s OpenAI-compatible API, getting started can be as simple as changing one line of code: swap the API URL to connect your existing app to our decentralized GPU network.
Same familiar request formats. No new SDK. No client rewrite.
That means less integration work and more time building what your users need.
From AI inference to fine-tuning, our goal is to make decentralized compute practical for the developers putting it to work.
Adoption starts with making the next step easier.
$CNX
YOU BUILT THE APP. CAN YOU AFFORD TO RUN IT?
A chatbot that answers questions. A tool that drafts replies. An assistant that writes captions.
Nothing extravagant. Just something useful that you want to put into people’s hands.
Then you look into the compute.
One provider has GPU quotas that could push back your launch. Another has a minimum rental commitment that stretches your budget before you’ve earned your first dollar.
You need enough compute to test an idea over a weekend. You’re looking at paying for a month.
For an independent developer, those costs can decide whether an app ever gets beyond a working prototype.
The gap between “I built this” and “I can afford to run this” is real.
Crynux is built to help close that gap.
Through an OpenAI-compatible API, developers can access AI inference on a decentralized GPU network and pay for the workloads they run, without taking on the cost of renting a dedicated GPU.
What would you build if compute costs stopped holding you back?
BITCOIN BROUGHT INDEPENDENT PARTICIPANTS TOGETHER TO POWER A MONETARY NETWORK. DECENTRALIZED AI BRINGS THAT SAME PRINCIPLE TO COMPUTE.
Blockchain showed us that independent participants could coordinate without a central intermediary. Decentralized AI applies that model to the infrastructure powering AI.
Instead of sourcing compute from a single provider, developers can access a network of independent GPU contributors around the world.
The opportunity goes beyond connecting GPUs. It means opening participation, reducing dependence on centralized infrastructure, and supporting AI workloads across a broader network of providers.
At Crynux, that’s the vision we’re building towards, through permissionless participation and verifiable execution.
Bitcoin powered a new way to transact without a central gatekeeper. Decentralized AI now challenges who gets to power intelligence.
Request ID for @coingecko
CL1310260027 geckoterminal geckoterminal.com/base/pools…
At Crynux, we're going beyond simply decentralizing GPU power.
We're building an AI compute network where execution is verifiable, infrastructure is permissionless, and the native token serves a real purpose.
But here's where Crynux becomes particularly interesting...
$CNX isn't just a token associated with the network. It's part of the infrastructure's economic foundation.
🔹 Compute payments - powering real AI inference and fine-tuning workloads.
🔹 Node staking - supporting reliable GPU compute and network security.
🔹 Delegation - allowing holders to support node operators and participate in rewards.
🔹 Governance - giving the community a role in shaping the network's future.
And this is why the utility matters. As more developers build on Crynux, more AI workloads can translate into greater network activity and potential demand for CNX.
Unlike tokens that depend entirely on market speculation, CNX has defined roles within a functioning decentralized AI compute ecosystem.
Verifiable compute.
Genuine utility.
Driven by real demand.
If you missed the latest CRYNUX AMA with @travladd, here’s a recap.
We explored the thinking behind Crynux, how the AI service works, and what participation looks like for developers, node operators and $CNX holders.
WHY CRYNUX EXISTS
The Crynux team began building in 2023 around a straightforward idea: a crypto network should do useful work. Developers pay for AI tasks, and compute providers earn by completing them. The ambition: an “AI version of Bitcoin”, a permissionless network powered by real demand for AI.
FROM GOOGLE TO DECENTRALIZED AI
The team’s experience at Google included work on image search, Android Nearby, on-device machine learning, TensorFlow Lite, federated learning and later LLMs. That background helped shape Crynux’s approach to distributing AI computation across devices.
AI THROUGH FAMILIAR TOOLS
The Crynux AI service provides an OpenAI-compatible API for accessing different AI models through familiar developer workflows. Payments are accepted in CNX, USDT and USDC, with free credits available to test the service in the playground.
PRICING THAT MATCHES THE JOB
Pricing responds to demand and task priority. Higher fees prioritize faster processing, while jobs that can tolerate a longer wait can use lower-cost options. One example discussed: leaving an AI coding agent working overnight instead of paying for an immediate response.
CONTRIBUTE COMPUTE OR API CAPACITY
Node operators run supported AI models and earn CNX for completed tasks. Compatible NVIDIA GPUs, MacBooks and Mac Minis can participate, depending on the model’s requirements. Providers can also contribute spare third-party API capacity so requests can be routed through it.
REWARDS TIED TO WORK
The tokenomics discussion centered on completed AI tasks, with approximately 71% of total supply allocated to task mining. We also covered Crynux’s fair launch, the team’s decision to reject VC proposals, and emissions spread over many years.
PARTICIPATION IN PRACTICE
Crynux DAO Convener, @Kaspa_HypeMan, joined the conversation and shared his experience running a node, seeing tasks arrive and having CNX holders delegate to it. His experience showed how node operators and token holders contribute to the network in practice.
BUILDING DEMAND
The team is working with developers to integrate Crynux into AI products. The closing message was direct: try the technology, inspect the code and judge the network by what it delivers.
You can listen to the full AMA here: nitter.cf/i/spaces/1mxPaZAOgyYKN
A great introduction to Crynux from one of our community members, showing how idle GPUs can power a decentralized AI network.
Watch the video for a quick look at the compute, verification and tokenomics behind Crynux.
We've already told you about @crynuxio with many posts.
For everyone who likes a video - here's one for you to know about a small cap AI gem we are heavily backing and working with 💪
✔️ Amazing team
✔️ 55 million tasks on testnet
✔️ On base and robinhood
→ An upcoming AMA with @travladd that you should join if you want to know more!
WE’RE LIVE IN 20 MINUTES
Join us and @travladd for a closer look at Crynux AI service, API keys and the upcoming marketplace.
Plus, a giveaway for listeners.
Tune in to hear what’s next for Crynux.
Last time I hosted base:0x9557dd9e241bc9636732623b672b4090af519396 was magical, don’t miss it this time, super solid project.
Giving away $200 to listeners.
Set a reminder: nitter.cf/i/spaces/1mxPaZAOgyYKN
NEW CRYNUX AMA WITH @TRAVLADD
Join us October 7th at 3 PM ET for a closer look at the new Crynux AI service, bringing decentralized GPU compute into familiar developer workflows.
We’ll also explore API keys, prepaid credits and the upcoming marketplace
Crynux DAO Convener @Kaspa_HypeMan will be joining the party as well.
See you there.
Put your GPU to work, or delegate $CNX to a node operator and share in its rewards.
Community guides like this make it easier to get started. Thanks for sharing your experience and helping others join the Crynux network.
Technical $CNX Crynux node & staking post!
With $CNX Crynux now on Robinhood and gaining traction, it's a great time to set up a node if you have a decent GPU & hardware or stake your balances to a node to support the network while earning weekly rewards — with no lockup on your initial staked balance. Setting up a node is simple on Windows, or staking your $CNX balance via Base or Robinhood is easy enough with the guide from @MiningRabid. If anyone needs any assistance setting up a node or staking please give me a DM or reach out to me as moderator on the official $CNX Telegram or Discord :) — Happy to help.
Here's a repost of @MiningRabid's node setup & staking guide:
youtube.com/watch?v=2Gfu0xMO…
As an example, on my own 5090 Windows node I've been receiving 20–30K $CNX in weekly rewards every Wednesday for completed tasks, with my 1M $CNX staked to the node.
Additionally, I provide a competitive 60% delegator reward/payout to anyone who stakes to my node, with emissions expected to be an estimated 26.1K+ $CNX every Wednesday to current delegators per the image below. Hop on & delegate if you like, or join another node if you don't have the hardware to run your own!:
portal.crynux.io/staking/nod…
portal.crynux.io/staking
Crynux $CNX is heading in the right direction. As the project progresses and handles more real-world AI workloads, we could see even more task activity and rewards across its decentralized compute network. 🦾🦾🚀🚀
@crynuxio @crynux_dao @RobinhoodApp @RobinhoodCrypto @base @safetrade @MiningRabid #crynux #nvidia $CNX
Curious how Crynux turns decentralized GPUs into verifiable AI infrastructure?
It starts with a permissionless network where distributed compute resources contribute to AI workloads such as inference, fine-tuning and training.
But decentralizing compute creates another challenge: how do you know the work was executed correctly?
That’s where vssML comes in.
Crynux uses verifiable random functions and zero-knowledge proofs to support secret sampling and validation. Selected tasks are independently executed by multiple nodes and their outputs compared, reducing the need to replicate every workload.
Crynux reports approximately 10% consensus overhead, making verification practical without sacrificing the benefits of decentralized compute.
The result is an AI compute network designed around three principles:
Permissionless participation.
Cryptographic verification.
Efficient scaling.
Decentralizing compute is one part of the equation. Making it verifiable, scalable and efficient is what completes it.
That’s Crynux.
CRYNUX IS LIVE ON ROBINHOOD CHAIN!
@crynuxio is the open AI compute network powered by idle consumer GPUs, supporting image generation, video generation, LLMs and fine-tuning, with computation verified through vssML. Mainnet went live in June.
$CNX is now bridged to @RobinhoodCrypto, expanding its reach to another ecosystem and creating more opportunities for people to discover and participate in Crynux.
The total supply remains fixed at 8,617,333,262 CNX, with 71% released only as real AI tasks are completed. Bridging CNX expands access without increasing supply.
Contract: 0x0A020A1B61b4315df6E1338BE93b7d986147729E
CRYNUX AMA RECAP WITH @travladd
From verified AI compute to the long-term vision for @crynuxio, here are the key takeaways:
VERIFICATION IS ESSENTIAL
Connecting GPUs is only part of the challenge. A decentralized network also needs to verify the work being performed, with economic safeguards that make cheating unattractive.
We discussed how Crynux approaches this through vssML, combining cryptographic proofs and secret sampling to verify computation without rerunning every task.
PERMISSIONLESS PARTICIPATION
GPU providers can download the node software and join without an approval process. Developers can access the network through an OpenAI-compatible API and build within familiar workflows.
You can contribute compute, build applications or delegate $CNX to a node operator and share in its rewards without running hardware yourself.
CNX AT THE CENTER OF THE NETWORK
CNX connects task payments, node rewards and staking. Its role is tied to the network’s activity: developers accessing compute, operators supplying it and participants helping secure the system.
The discussion explored how broader adoption and more AI workloads could expand the use of CNX throughout the ecosystem.
COMPUTE THAT FITS THE WORKLOAD
A chatbot needs a fast response; a background research task can wait.
We explored how task fees determine queue priority, giving developers flexibility to balance speed and cost around their application’s needs.
LOOKING AHEAD: AI AND WEB3
The conversation also explored longer-term possibilities beyond compute, including model tokenization, cross-chain model assets and AI-Fi applications.
A strong sense of optimism emerged around how these possibilities could extend the CNX ecosystem from compute infrastructure into a broader network of AI services and assets.
Verified compute provides the foundation for a broader ecosystem where developers can build, deploy and bring new AI services to the market.
EXPERIENCE BEHIND THE VISION
The AMA also touched on the experience behind our work: years across AI infrastructure, blockchain and cryptography, including work at Google and research in computation verification and privacy-preserving machine learning.
Thank you to @travladd, @BraverCrypto and everyone who joined the conversation.
Catch the full recording below.
If you’ve never heard of @crynuxio jump in for this one🔥👀
$200 will also be given away live.
Set a reminder: nitter.cf/i/spaces/1qJVmyOXBeMGB