Sentient has unveiled $SENT tokenomics ahead of its sale
Sentient is a decentralized, open-source AGI project aimed at building community-owned AI infrastructure.
The SENT token will have a total supply of 34.35B tokens, with 44% allocated to airdrops, community grants, bounties, and incentives.
blog.sentient.xyz/posts/tokenomics
Sentient is a decentralized, open-source AGI project aimed at building community-owned AI infrastructure.
The SENT token will have a total supply of 34.35B tokens, with 44% allocated to airdrops, community grants, bounties, and incentives.
blog.sentient.xyz/posts/tokenomics
Q1 2026 mainnet season is stacked.
Tempo ($500M), MegaETH ($108M), Gensyn ($67M), Nexus ($27M), Miden ($25M), Pod ($10M), Pharos ($8M), and Idos ($6.5M) are all going live.
That’s billions in capital finally moving from fundraising to real products.
This is where narratives end and real adoption starts.
Tempo ($500M), MegaETH ($108M), Gensyn ($67M), Nexus ($27M), Miden ($25M), Pod ($10M), Pharos ($8M), and Idos ($6.5M) are all going live.
That’s billions in capital finally moving from fundraising to real products.
This is where narratives end and real adoption starts.
🚀 @SentientAGI is becoming one of the strongest examples that open source AI can compete with closed models like GPT-5 and Gemini in real benchmarks and real usage.
ROMA and SERA show two different parts of this clearly.
ROMA is focused on complex reasoning.
On the SEALQA benchmark, ROMA scored 45.6%, which is higher than:
- Kimi Researcher at 36%
- Gemini 2.5 Pro at 19.8%
- And Most open source multi agent systems
This is not a small gap. It shows ROMA is solving structured, multi-step problems better than many large proprietary systems. It also runs fully open source, with full task tracing and tool execution visibility.
SERA is focused on real time crypto and Web3 research.
On the DMind benchmark, it ranks #1 among open source agents and is only 2–3% behind GPT-5 Medium.
It also beats GPT-5, Grok 4, Gemini, and Perplexity Finance on Sentient’s internal crypto benchmark.
What makes SERA different is how it works:
- It uses 50+ live APIs like CoinGecko, DefiLlama, Arkham, and on chain data
- Data freshness is 24–48 hours, compared to 7+ days for general models
- Latency is under 45 seconds
- It avoids LLM randomness for tool routing, which makes results more consistent
- It runs on fully open models like GPT-OSS-120B and Qwen-Next-80B
This is why SERA performs well in production style research, not just demos.
Now look at scale:
290,000+ users
22 million+ queries processed
Around 13 queries per second
110+ GRID partners across agents, data providers, and models
That means this system is already handling real traffic, not just lab testing.
Sentient is not building a single AI product. It is building an AI network.
Users pay agents in $SENT.
Agents pay data providers in SENT.
Builders earn revenue based on usage and quality.
Stakers secure the system and govern it.
That is a full economic loop around intelligence.
ROMA and SERA show two different parts of this clearly.
ROMA is focused on complex reasoning.
On the SEALQA benchmark, ROMA scored 45.6%, which is higher than:
- Kimi Researcher at 36%
- Gemini 2.5 Pro at 19.8%
- And Most open source multi agent systems
This is not a small gap. It shows ROMA is solving structured, multi-step problems better than many large proprietary systems. It also runs fully open source, with full task tracing and tool execution visibility.
SERA is focused on real time crypto and Web3 research.
On the DMind benchmark, it ranks #1 among open source agents and is only 2–3% behind GPT-5 Medium.
It also beats GPT-5, Grok 4, Gemini, and Perplexity Finance on Sentient’s internal crypto benchmark.
What makes SERA different is how it works:
- It uses 50+ live APIs like CoinGecko, DefiLlama, Arkham, and on chain data
- Data freshness is 24–48 hours, compared to 7+ days for general models
- Latency is under 45 seconds
- It avoids LLM randomness for tool routing, which makes results more consistent
- It runs on fully open models like GPT-OSS-120B and Qwen-Next-80B
This is why SERA performs well in production style research, not just demos.
Now look at scale:
290,000+ users
22 million+ queries processed
Around 13 queries per second
110+ GRID partners across agents, data providers, and models
That means this system is already handling real traffic, not just lab testing.
Sentient is not building a single AI product. It is building an AI network.
Users pay agents in $SENT.
Agents pay data providers in SENT.
Builders earn revenue based on usage and quality.
Stakers secure the system and govern it.
That is a full economic loop around intelligence.