BTC $77,669.23 +0.23%
ETH $2,257.21 +0.24%
BNB $707.33 +0.14%
XRP $1.24 -0.86%
SOL $93.95 -1.97%
TRX $0.3126 +1.83%
DOGE $0.0750 -1.07%
ADA $0.1910 -1.20%
BCH $280.44 -2.37%
LINK $11.40 -3.78%
HYPE $87.06 -0.84%
AAVE $133.38 -0.75%
SUI $0.8887 -1.27%
XLM $0.1669 -1.56%
ZEC $1,221.77 -4.37%
AAPL $344.36 -0.14%
AMZN $241.28 +0.24%
GOOGL $283.67 +0.22%
MSFT $525.26 +0.16%
META $614.71 +0.10%
NVDA $204.69 +0.29%
TSLA $399.70 +0.07%
SNDK $1,160.82 +0.68%
INTC $65.92 +1.54%
SPCX $137.61 +0.07%
MU $900.94 +0.29%
AMD $509.34 +0.27%
BTC $77,669.23 +0.23%
ETH $2,257.21 +0.24%
BNB $707.33 +0.14%
XRP $1.24 -0.86%
SOL $93.95 -1.97%
TRX $0.3126 +1.83%
DOGE $0.0750 -1.07%
ADA $0.1910 -1.20%
BCH $280.44 -2.37%
LINK $11.40 -3.78%
HYPE $87.06 -0.84%
AAVE $133.38 -0.75%
SUI $0.8887 -1.27%
XLM $0.1669 -1.56%
ZEC $1,221.77 -4.37%
AAPL $344.36 -0.14%
AMZN $241.28 +0.24%
GOOGL $283.67 +0.22%
MSFT $525.26 +0.16%
META $614.71 +0.10%
NVDA $204.69 +0.29%
TSLA $399.70 +0.07%
SNDK $1,160.82 +0.68%
INTC $65.92 +1.54%
SPCX $137.61 +0.07%
MU $900.94 +0.29%
AMD $509.34 +0.27%

ben

All
Article
Flash

Grok 4.8 has not been released yet, and Musk said he doesn't care about the benchmark scores

On October 30, Musk admitted that the development of Grok 4.8 took longer than expected but assured that the new model would not disappoint. He stated that the team does not optimize for benchmark testing and he himself does not care about scores, focusing more on the practical applications of the model. On September 14, Musk revealed that Grok 4.8 has 2.5 trillion parameters and uses a brand new C++ training system, expecting to complete pre-training by the end of that week, followed by reinforcement learning.Now, more than six weeks have passed, and the new model has still not been officially released. The independent evaluation agency Artificial Analysis found that under the same xhigh inference setting, Grok 4.7's comprehensive intelligence index only increased by 2 points compared to Grok 4.6, rising from 44 to 46 points, while the average output tokens per question increased from about 38,000 to 81,000. Users in the Cursor community have also complained that Grok 4.7 consumes resources faster, but the actual improvement is not significant.In an interview with CCTV in September, Musk admitted that Grok 4.7 is not as good as Claude Opus 5.5, predicting that it would not catch up to the top tier until 2027. In October, he proposed allowing Grok Bot to call external models, including Claude, based on tasks, and plans to have Grok 4.8's high-speed version handle a large number of simple requests.

first_img Celestia launched the payment architecture Voyager, with benchmark tests reaching 21 million TPS

The modular blockchain network Celestia announced the launch of the application-specific blockchain architecture Voyager, aimed at large-scale payment scenarios in agent commerce. The official statement claims that Voyager processed approximately 21 million transactions per second continuously between 100 million accounts during its first complete benchmark test, which lasted for 5 hours.Celestia stated that Voyager splits payment processing into independently scalable threads, avoiding blocking the execution thread responsible for updating balances. Transactions are first verified for signatures and packaged at the entry layer, then enter a circular buffer inspired by the exchange system Disruptor; execution logic is built-in, and state is stored in a flat table rather than a Merkle tree. Each transaction is published to Celestia's data availability protocol Fibre, for external replay and verification of whether payments were executed correctly.The benchmark test covered approximately 40 billion random transfers, with an average throughput of 21.2 million TPS and a peak of 21.6 million TPS. Execution ran on a single CPU core, with node memory around 48 GB. This test was conducted on a single node and did not include state root computation, cross-machine replication, or real-time state queries at the same scale. Celestia claims it is advancing Voyager into production and will release more in-depth technical documentation in the coming weeks. Fibre was launched on the mainnet last week, and Voyager is the first execution layer product launched after its acquisition of Sovereign Labs.
app_icon
ChainCatcher Building the Web3 world with innovations.