SumPlus @SumPlusReal

Jev is a strong signal for where AI apps are heading. ⚡ If AI can decide UI structure, actions, and data flow faster than traditional model-driven rendering, the app layer becomes much more dynamic. At Sumplus, we’re excited about this direction: SafeRouter for trusted model access. Arsenal for callable skills. Jev for real-time, task-native interfaces. #jev @typesafeai

OpenRouter @OpenRouter

2 days ago

1/ Jev, a decision model by @typesafeai, sparked a burst of projects and discussion. We tested it using Ori Eval against popular LLMs on OpenRouter at judging. Jev was >5x faster than the next fastest model, and even its slowest requests beat every other model's median.

72 61 947 55K 223

Jev is changing how AI apps are built. ⚡ Instead of fixed interfaces, AI apps can now assemble components, actions, and data in real time. That’s exactly where Sumplus fits in. With SafeRouter, agents get trusted model access. With Arsenal, agents get callable skills and tools. With #JeV @typesafeai , those capabilities can turn into dynamic, task-native interfaces. The future app won’t be static. It will be generated around the user, the task, and the agent workflow. 🚀 #ai #JeV

80% of Robinhood Stock Token volume trading on Uniswap is a major signal. ⚡ Tokenized equities are not just moving onchain — they are becoming part of DeFi liquidity. The next step is AI. @Uniswap AI agents will monitor these markets, compare liquidity, manage risk, and execute strategies across DeFi rails in real time. At Sumplus, we’re building for that future: DeFi liquidity + tokenized assets + AI execution. 🚀 #uni #robinhood

Uniswap @Uniswap

3 days ago

80% of all Robinhood Stock Token volume is traded on Uniswap That translates to $10B (and counting)

129 149 1K 79K 43

@arc done

Build on Arc.

Arc mainnet is live. ⚡ A stablecoin-native economic OS with USDC gas, sub-second finality, EVM compatibility, and infrastructure for payments, trading, tokenized assets, and agentic workflows. For AI agents, this is a major step forward. Agents need financial rails they can use in real time — and trusted model infrastructure to think, route, and execute safely. @arc That’s what SafeRouter is built for: Private model access. Verifiable routing. Agent-ready infrastructure. Arc brings the economic rails. SafeRouter brings trusted AI access. 🚀 #ai #arc #agent

Arc @arc

5 days ago

Arc Mainnet is live. Arc launches as the Economic OS for the internet: an open platform for global markets, real-time value movement, tokenized assets, and agentic economic activity. Arc is more than a blockchain. It launches as a full-stack financial platform with assets,

2K 2K 8K 2.8M 580

The minting of ARC marks an important technical milestone for the Arc roadmap. ⚡ Stablecoin-native infrastructure is moving from concept to execution — and this is exactly where AI agents will need trusted rails. At @SumPlusReal , we’re building for that future: AI agents, onchain payments, verifiable execution, and safer model access. Arc is building the financial rails. Sumplus is building the agent execution layer. 🚀 #ai #ARC

Arc @arc

5 days ago

The ARC token has been minted. Circle is the first publicly traded company to mint a network token for a new blockchain. This is a technical milestone in the Arc roadmap as the network explores a future transition from Proof of Authority toward Proof of Stake. The minting of

734 621 4K 1.3M 537

See you soon AI agents need trusted models, useful tools, and secure execution. That’s what we’re building at Sumplus with SafeRouter, Arsenal, and Maria. Excited for Arc’s next chapter. @arc 🫡 #ai

Arc @arc

5 days ago

Calling all Architects. See you soon.

2K 775 7K 568K 117

@arc See you soon, @arc 🫡 AI agents need trusted models, useful tools, and secure execution. That’s what we’re building at Sumplus with SafeRouter, Arsenal, and Maria. Excited for Arc’s next chapter.

DeepSeek V4.1-Flash is becoming a serious choice for coding agents. ⚡ 74.34% pass@1 on DeepSWE — around GPT-6 Astra level, but at a fraction of the cost. For agent workflows, lower cost means more retries, more loops, and more chances to get the right result. At @SumPlusReal SafeRouter, we make DeepSeek easier to access with privacy, verifiable routing, and developer-friendly API access. Cheaper no longer means weaker. 🚀 #ai #LLM #agent

Arc @arc

5 days ago

Calling all Architects. See you soon.

2K 775 7K 568K 117

DeepSeek V4.1-Flash is changing the cost equation for coding agents. ⚡ According to recent benchmark results, it reached 74.34% pass@1 on DeepSWE, slightly ahead of GPT-6 Astra — while costing around 15x less per task. It’s becoming the practical choice for real agent workloads. 🚀

DeepSeek-V4.1-Flash keeps proving its value. ⚡ Ranked #3 among open models in Agent Arena, with a median cost of only $0.07 per task. That’s close to top open-model performance, but at a much lower cost. For builders and AI agents, this is the real shift: stronger models, lower costs, more practical agent workflows. At @SumPlusReal SafeRouter, we’re making DeepSeek models easier to access with privacy protection, verifiable routing, and developer-friendly API access. Cost-efficient intelligence is becoming infrastructure. 🚀 #ai #sumplus #agent

Arena.ai @arena

7 days ago

Exciting news: DeepSeek-V4.1-Flash (Max) by @deepseek_ai just landed in Agent Arena at #3 among open models! With +4.87% net improvement and a median cost per task of $0.07 it reshaped the Pareto frontier. Among the top 3 open models, DeepSeek-V4.1-Flash (Max) has the lowest

38 57 941 294K 116

Kimi is becoming hard to ignore. ⚡ The gap between top closed models and Chinese open-weight models is getting smaller fast. At @SumPlusReal SafeRouter, we already support Kimi — giving builders and AI agents easier access to frontier Chinese models with privacy protection, verifiable routing, and developer-friendly API access. More models. More choice. One trusted gateway. 🚀 #ai #kimi #glm

Bhavy☄️ @Bhavani_00007

a week ago

GPT-6 Astra vs Kimi K3. I'm comparing these because the gap between US models and Chinese models is already getting small. If US AI labs slow down, they will be the ones most affected by it. > Astra did it in one shot 6m 20s and $8. > Kimi K3 got almost the same result, but I

31 6 200 26K 31

Kimi is becoming hard to ignore. ⚡ The gap between top closed models and Chinese open-weight models is getting smaller fast. At Sumplus SafeRouter, we already support Kimi — giving builders and AI agents easier access to frontier Chinese models with privacy protection, verifiable routing, and developer-friendly API access. More models. More choice. One trusted gateway. 🚀

Sumplus has fully integrated Arc. ⚡ As @arc goes live on Sept. 16, we’re ready to support the next wave of stablecoin-native applications, onchain payments, and AI agent workflows. With SafeRouter, Arsenal, and Maria, Sumplus is building trusted execution infrastructure for the agentic finance era. Arc brings stablecoin-native rails. Sumplus brings AI agent execution. More soon. 🚀 #ai #sumplus #arc

Code Arena shows how fast the model layer is evolving. ⚡ In one year, frontier WebDev performance jumped by +340 pts, with more labs competing at the top — OpenAI, Anthropic, Qwen, Kimi and others. For builders, this means one thing: There won’t be one model for every task. AI agents will need reliable infrastructure to access, compare, and route across frontier models in real time. That’s what Sumplus SafeRouter is building for. 🚀 #ai #sumplus

Arena.ai @arena

a week ago

Nearly a year of Code Arena: WebDev progress compressed into 15 seconds. Each line follows the highest-scoring model from top labs over time, showing the pace of improvement across the ecosystem. In just the last year, the model in the leading spot increased score by +340 pts

25 47 626 93K 117

DeepSeek-V4.1-Flash is showing strong momentum in Code Arena WebDev. ⚡ ~#14 overall with 1620 pts, and ~#4 among open models — while priced at only $0.30 / $1.20 per 1M tokens. Better performance. Much lower cost. For builders and AI agents, this is exactly why DeepSeek is becoming one of the most cost-efficient model choices today. At @SumPlusReal SafeRouter, we’re making DeepSeek models easier to access with official model access, privacy protection, and developer-friendly pricing. Open models are getting stronger. The infra layer needs to keep up. 🚀 #ai #deepseek

Arena.ai @arena

2 weeks ago

DeepSeek-V4.1-Flash by @deepseek_ai just landed ~#14 overall in Code Arena: WebDev with 1620 pts (AutoEval)! Among open models, DeepSeek-V4.1-Flash landed at ~#4 within 11 pts of Qwen3.8-Flash-Next. This release is a significant improvement compared to DeepSeek-V4 variants: +38

43 78 1K 153K 142

DeepSeek-V4.1-Flash looks like a major step forward. ⚡ Smarter, faster, more efficient — and now with native visual understanding. At Sumplus SafeRouter, we’re excited to support the next wave of @deepseek_ai models with trusted access, privacy protection, and developer-friendly pricing. Better models need better infrastructure. 🚀 #ai #deepseek

DeepSeek @deepseek_ai

2 weeks ago

🚀 Introducing DeepSeek-V4.1-Flash: smarter, faster, more efficient. 🔹 Introducing the smallest model in our new architecture family, with native visual understanding. 🔹 Designed for greater capability, faster inference, higher throughput, and scaling to larger models. 1/6

1K 3K 28K 6.8M 5K

DeepSeek V4.1 Flash is becoming hard to ignore. ⚡ 98% of GPT-6 Astra’s score at 1.4% of the cost is not just a benchmark result — it’s a market signal. Open models are getting stronger, faster, and far more cost-efficient. At @SumPlusReal SafeRouter, we’re making DeepSeek models easier to access with official model access, privacy protection, and developer-friendly pricing. More performance. Lower cost. Better access for builders. 🚀 #ai #sumplus

OpenDesign @OpenDesignHQ

2 weeks ago

We benchmarked DeepSeek V4.1 Flash by @deepseek_ai . It reached 98% of GPT-6 Astra’s score at 1.4% of the cost on everyday design tasks based on user requests. Every model except Astra scored lower AND cost more. Are open models overtaking closed ones? Full results below ↘️

329 991 11K 2.8M 4K

GLM-5.3 is looking strong for agent and coding workloads. ⚡ At Sumplus SafeRouter, we’re preparing to bring GLM-5.3 online soon — with privacy protection, verifiable routing, and developer-friendly API access. @ArtificialAnlys More frontier models. One trusted gateway. 🚀 #Glm

Artificial Analysis @ArtificialAnlys

a month ago

GLM-5.3 achieves 60 on the Artificial Analysis Intelligence Index, on par with Kimi K3 and up 7 points from GLM-5.2. Once the weights are released it will be tied as the leading open weights model @Zai_org has just launched GLM-5.3, which ties Kimi K3 (60) for the most

94 155 2K 233K 255