NVIDIA-backed Reflection AI releases Beam model with 501 billion parameters, claiming significantly lower inference computation costs than Chinese competitor GLM-5.2.
NVIDIA-backed Reflection AI has officially released its first open-weight AI model, Beam. The company stated that Beam's performance in advanced reasoning benchmarks is comparable to China's leading open model, GLM-5.2, but with a 3-4x reduction in inference computational cost. Beam is a model with 501 billion parameters, of which 23 billion are active parameters, pre-trained on 23.8 trillion tokens of data, and has a context window of 1 million tokens. Reflection AI has raised approximately $4.7 billion from investors including NVIDIA and Sequoia Capital, with the company valued at $25 billion. Beam's weights and full technical details are expected to be released this month.
No AI analysis yet. Tap the "AI Analysis" button above to generate one now.
Disclaimer: This content reflects the author's personal views only and does not constitute investment advice.
24H Trending
-
1
Regulatory Landscape and Exchange Policies for Virtual Currencies in Mainland China
-
2
Cerebras's pre-market stock rose 6% after OpenAI CEO Altman called Cerebras a "close partner."
-
3
Bitcoin (BTC): Latest Updates and Market Analysis of Digital Gold
-
4
Litecoin's price has recently surged: Can ETF applications and halving expectations sustain its rally?
-
5
Cryptocurrency Information Platforms and Software Download Guide: Secure Access to Industry News and Trading Tools
-
6
Multiple U.S. states will vote on wealth taxes in November, with California potentially imposing a 5% net worth tax on billionaires.
-
7
Ethereum: Market Impact, Technological Advancements, and Future Outlook
-
8
Decoding the Connection Mechanism of Bitcoin Private Keys, Public Keys, and Wallet Addresses
-
9
STEEM Coin Analysis: The Current Status and Development Challenges of a Decentralized Social Blockchain
-
10
SOL Rebounds Strongly: Can Solana Reach New All-Time Highs?
Markets Today
Recommended Reading










