Cerebras Beats NVIDIA Blackwell in Llama 4 Maverick Inference
'Cerebras has beaten the Llama 4 Maverick inference speed record set by NVIDIA last week. Artificial Analysis benchmarked Cerebras' Llama 4 Maverick endpoint at 2,522 t/s compared to NVIDIA Blackwell's 1,038 t/s for the same model." - Artificial Analysis
Share
'Cerebras has beaten the Llama 4 Maverick inference speed record set by NVIDIA last week,' said Micah Hill-Smith, Co-Founder and CEO of Artificial Analysis. 'Artificial Analysis has benchmarked Cerebras' Llama 4 Maverick endpoint at 2,522 tokens per second, compared to NVIDIA Blackwell's 1,038 tokens per second for the same model. We've tested dozens of vendors, and Cerebras is the only inference solution that outperforms Blackwell for Meta's flagship model.'
With today's results, Cerebras has set a world record for LLM inference speed on the 400B parameter Llama 4 Maverick model, the largest and most powerful in the Llama 4 family. Artificial Analysis tested multiple other vendors, and the results were as follows: SambaNova 794 t/s, Amazon 290 t/s, Groq 549 t/s, Google 125 t/s, and Microsoft Azure 54 t/s.
Andrew Feldman, CEO of Cerebras Systems, said, 'The most important AI applications being deployed in enterprise today—agents, code generation, and complex reasoning—are bottlenecked by inference latency. These use cases often involve multi-step chains of thought or large-scale retrieval and planning, with generation speeds as low as 100 tokens per second on GPUs, causing wait times of minutes and making production deployment impractical. Cerebras has led the charge in redefining inference performance across models like Llama, DeepSeek, and Qwen, regularly delivering over 2,500 TPS/user.'
With its world record performance, Cerebras is the optimal solution for Llama 4 in any deployment scenario. Not only is Cerebras Inference the first and only API to break the 2,500 TPS/user milestone on this model, but unlike the Nvidia Blackwell used in the Artificial Analysis benchmark, the Cerebras hardware and API are available now. Nvidia used custom software optimizations that are not available to most users. Interestingly, none of the Nvidia's inference providers offer a service at Nvidia's published performance. This suggests that in order to achieve 1000 TPS/user, Nvidia was forced to reduce throughput by going to batch size 1 or 2, leaving the GPUs at less than 1% utilization. Cerebras, on the other hand, achieved this record-breaking performance without any special kernel optimizations, and it will be available to everyone through Meta's API service coming soon.
For cutting-edge AI applications such as reasoning, voice, and agentic workflows, speed is paramount. These AI applications gain intelligence by processing more tokens during the inference process. This can also make them slow and force customers to wait. And when customers are forced to wait, they leave and go to competitors who provide answers faster—a finding Google showed with search more than a decade ago.
With record-breaking performance, Cerebras hardware and resulting API service is the best choice for developers and enterprise AI users around the world.
For more information, please visit https://www.cerebras.ai/.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Business Insider
an hour ago
- Business Insider
Top Analysts Boost Meta Platforms Stock Price Target Ahead of Q2 Earnings
Top analysts from Jefferies and Canaccord Genuity boosted their price targets for Meta Platforms (META) stock ahead of the social media giant's Q2 earnings on July 30. While Jefferies' 5-star analyst Brent Thill raised the price target for META stock from $790 to $845, Canaccord Genuity's top analyst Maria Ripps increased her price target from $825 to $850. Both top-rated analysts reaffirmed a Buy rating on Meta Platforms stock, reinforcing their confidence in the company's growth potential. Elevate Your Investing Strategy: Take advantage of TipRanks Premium at 50% off! Unlock powerful investing tools, advanced data, and expert analyst insights to help you invest with confidence. Make smarter investment decisions with TipRanks' Smart Investor Picks, delivered to your inbox every week. These price target hikes came in even as META CEO Mark Zuckerberg faces an $8 billion shareholder lawsuit over accusations of privacy breach. Meanwhile, Wall Street expects Meta Platforms to report earnings per share (EPS) of $5.84, reflecting a 13.2% year-over-year growth. Top Jefferies Analyst Is Upbeat About Meta's Q2 Earnings Thill raised his Q2 and full-year revenue estimates for Meta Platforms by 3.1% and 2.6%, respectively. The 5-star analyst expects Meta to deliver revenue of $45.2 billion in Q2, suggesting a 15.7% year-over-year growth, which is higher than the Street's consensus estimate of 14.2%. Additionally, Thill contends that Meta Platforms' Q3 guidance seems achievable, given easier comparisons and some conservatism in the Street's estimates. Meanwhile, Thill believes that Meta's $14.3 billion investment in Scale AI and the appointment of Alexandr Wang, founder of Scale AI, as the Chief AI Officer, along with the hiring of many other high-profile researchers, indicate the company's intention to streamline its AI organization and revamp leadership following the disappointing response to Llama 4. Thill argues that while Meta's capex is expected to stay elevated and put pressure on its near-term earnings, he remains confident about the long-term return on investment (ROI). Thill explained that his constructive view is supported by favorable checks on CPM (cost per mille), an increase in time spent as indicated by SensorTower, and other positive metrics. CPM indicates the cost an advertiser pays for one thousand impressions of an ad. Canaccord Views META Stock as a Top Digital Ad Pick Ripps stated that META stock remains Canaccord Genuity's top digital advertising pick. Despite its premium valuation, Ripps continues to like META stock, driven by several tailwinds. The analyst expects the company to report impressive Q2 results, backed by mid-teens year-over-year growth in ad revenue. For Q2, Ripps expects both ad revenue and total revenue to grow by about 14% year-over-year, with the modest sequential deceleration reflecting tariff-related uncertainty. The analyst expects Meta's growth to be driven by continued AI-driven improvements to content creation and ad recommendation models. Notably, the company launched a new generative ad recommendation model in Q1 2025, which is twice as efficient at improving ad performance as legacy models. Looking ahead, Ripps expects the pace of innovation to remain robust at Meta, bolstered by the acquisition of a 49% stake in Scale AI, hiring of OpenAI and Apple (AAPL) researchers, unveiling of Meta Superintelligence labs, and the acquisition of voice AI startup PlayAI. Ripps noted that while META stock is trading near all-time highs, the setup continues to look attractive, particularly as we move into 2026. Is META a Good Stock to Buy? Overall, Wall Street is bullish on Meta Platforms stock, with a Strong Buy consensus rating based on 41 Buys and four Hold recommendations. The average META stock price target of $737.86 indicates a 5% upside potential. META stock has risen 20% year-to-date.
Business Wire
2 hours ago
- Business Wire
BOWS Attorney Ross Smith Earns Lawdragon Next Generation Honors
HOUSTON--(BUSINESS WIRE)--Bissinger, Oshman, Williams & Strasburger LLP (BOWS) attorney Ross Smith has earned recognition on the 2025 Lawdragon 500 X – The Next Generation list. Honoring attorneys who are in their first 15 years of practice, The Next Generation guide offers a 'forecast of the fascinating future of global law practice,' according to Lawdragon, the publisher of more than two dozen guides focused on the legal world's changemakers. Mr. Smith represents corporate and individual clients in lawsuits and disputes involving breach of contract claims, breach of fiduciary duty claims, fraud claims, construction defect claims, trade secret claims, non-compete issues, insurance claims, municipal issues and securities claims. 'This is a well-deserved honor,' says BOWS co-founder David Bissinger. 'It has been exciting to watch Ross grow and develop into a truly exceptional litigator.' This is the second consecutive year Mr. Smith has earned a place on Lawdragon's 500 X – The Next Generation list. He also previously earned Lawdragon 500 Leading Litigators in America and Best Lawyers: Ones to Watch recognition. Bissinger, Oshman, Williams & Strasburger LLP is a Houston-based business trial and transaction firm focused on providing impactful, cost-effective solutions to complex disputes and transactions requiring careful attention, extensive experience and a high level of sophistication.
Yahoo
3 hours ago
- Yahoo
The Stock Market Just Did Something for the 11th Time Since 1984. History Says It Signals a Big Move in the Next Year.
Key Points The S&P 500 outperformed the S&P 500 Equal Weight Index by more than a percentage point in the first half of 2025, something it has done just 11 times since 1984. Following incidents where the S&P 500 beat its equal-weight counterpart in the first half of a given year, the S&P 500 has returned an average of 21% in the next 12 months. The S&P 500 currently trades at 22.3 times forward earnings, an unusually expensive valuation that has historically correlated with a three-year return of just 3% annually. 10 stocks we like better than S&P 500 Index › The S&P 500 (SNPINDEX: ^GSPC) added 5.5% in the first half of 2025 as the economy remained strong despite sweeping tariffs from the Trump administration. The "Magnificent Seven" stocks were responsible for 15% of those gains because of particularly strong performances from Meta Platforms, Microsoft, and Nvidia. Put differently, a relatively small number of mega-cap companies accounted for a relatively large portion of the S&P 500's gains in the first half of the year. In fact, the S&P 500 beat the S&P 500 Equal Weight Index -- an index that affords each stock the same weight, rather than weighting them by market capitalization -- by 1.7 percentage points. So what? That was the 11th time since 1984 the S&P 500 has outperformed its equal-weight counterpart by more than a percentage point in the first half of the year. The last 10 times, the S&P 500 usually rocketed higher in the subsequent 12 months as the upward momentum broadened throughout the index. Here's what investors should know. History says the S&P 500 could rocket higher in the next year The S&P 500 tracks 500 large-cap stocks. The index is weighted by market value, meaning more valuable companies have a greater impact on its performance. Conversely, the S&P 500 Equal Weight Index tracks the same stocks, but places the same weight on each one, so no company impacts its performance more than any other. Since 1984, the S&P 500 has beat its equal-weight counterpart by more than a percentage point during the first half of 11 years , with the most recent incident being 2025. The following table shows the 12-month return in the S&P 500 after the last 10 incidents. Interestingly, the index has always increased and almost always achieved double-digit returns. Year S&P 500 Forward 12-Month Return 1984 25% 1990 4% 1995 23% 1997 28% 1998 21% 2012 18% 2017 12% 2020 39% 2023 23% 2024 14% Average 21% Data source: YCharts. Chart by Author. As shown, when the S&P 500 beats its equal-weight counterpart by over a percentage point in the first half of the year, it has returned an average of 21% in the next 12 months. Past performance is never a guarantee of future results, but we can use that data to make an educated guess about what comes next for the stock market. The S&P 500 closed at 6,205 on June 30, 2025. The index will increase 21% to 7,508 by June 30, 2026, if its performance matches the historical average. That implies 20% upside from its current level of 6,244. The S&P 500 currently trades at a historically expensive valuation One reason the S&P 500 beat its equal-weight counterpart in the first half of 2025 was strong earnings growth from the largest companies. The "Magnificent Seven" in aggregate saw earnings increase 28%, while the other 493 stocks in the index saw earnings increase just 9%, according to FactSet Research. Importantly, Wall Street analysts generally think the gap will narrow in the future, as follows: The "Magnificent Seven" are expected to report 16% earnings growth in 2025, while the other 493 companies in the S&P 500 are forecast to report 7% earnings growth. The "Magnificent Seven" are expected to report 15% earnings growth in 2026, while the other 493 companies in the S&P 500 are forecast to report 13% earnings growth. Despite being such large companies, the "Magnificent Seven" are still reporting strong earnings. That bodes well for the S&P 500 because the "Magnificent Seven" comprise one-third of the index by market value. But it's even more encouraging to know earnings are forecast to accelerate across the other 493 companies. That suggets the stock market rally could broaden in the coming months. However, the S&P 500 currently trades at 22.3 times forward earnings, a premium to the 10-year average of 18.4 times forward earnings, according to FactSet Research. Historically, valuations near 22 times forward earnings have correlated with annual returns of just 3% over the next three years, according to economist Torsten Slok at Apollo Global Management. Here's the big picture: History says the S&P 500 could rocket higher in the next year. But history also says elevated valuations could lead to weak results over the next three years. Investors can reconcile those opposing views by limiting purchases to high-conviction stocks that trade at reasonable valuations. Do the experts think S&P 500 Index is a buy right now? The Motley Fool's expert analyst team, drawing on years of investing experience and deep analysis of thousands of stocks, leverages our proprietary Moneyball AI investing database to uncover top opportunities. They've just revealed their to buy now — did S&P 500 Index make the list? When our Stock Advisor analyst team has a stock recommendation, it can pay to listen. After all, Stock Advisor's total average return is up 1,060% vs. just 179% for the S&P — that is beating the market by 881.02%!* Imagine if you were a Stock Advisor member when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you'd have $679,653!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you'd have $1,046,308!* The 10 stocks that made the cut could produce monster returns in the coming years. Don't miss out on the latest top 10 list, available when you join Stock Advisor. See the 10 stocks » *Stock Advisor returns as of July 15, 2025 Randi Zuckerberg, a former director of market development and spokeswoman for Facebook and sister to Meta Platforms CEO Mark Zuckerberg, is a member of The Motley Fool's board of directors. Trevor Jennewine has positions in Nvidia. The Motley Fool has positions in and recommends FactSet Research Systems, Meta Platforms, Microsoft, and Nvidia. The Motley Fool recommends the following options: long January 2026 $395 calls on Microsoft and short January 2026 $405 calls on Microsoft. The Motley Fool has a disclosure policy. The Stock Market Just Did Something for the 11th Time Since 1984. History Says It Signals a Big Move in the Next Year. was originally published by The Motley Fool 擷取數據時發生錯誤 登入存取你的投資組合 擷取數據時發生錯誤 擷取數據時發生錯誤 擷取數據時發生錯誤 擷取數據時發生錯誤
