SemiAnalysis: Kimi K3 Ranks Third Globally, Could Reveal Hidden Profit Margins of OpenAI and Anthropic
SemiAnalysis analysts Jordan Nanos and Max Kan recently analyzed Kimi K3, the model developed by Chinese AI startup Moonshot AI, concluding that it outperforms Google Gemini in comprehensive benchmark tests. This not only reflects the narrowing gap between Chinese and U.S. AI models but also offers new insights into the business models of closed-source AI firms like Anthropic and OpenAI.
According to SemiAnalysis’s overall assessment, Kimi K3 currently ranks third globally, trailing only Fable 5 and GPT-5.6, and surpassing Google Gemini. The analysts noted that while this result does not signal major issues for Google’s AI business, Kimi K3’s publicly disclosed parameter count, performance, and pricing provide a reference for external estimates of the economic value of closed-source models.
Kimi K3 has 2.8 trillion parameters, far exceeding most open-source models. Jordan Nanos stated that a model of this size cannot be deployed on a single NVIDIA B200 GPU, requiring higher-spec hardware such as GB300, B300-class systems, or AMD MI355X. Based on this, he speculated that Anthropic and OpenAI’s flagship closed-source models likely operate at a similar parameter scale, rather than holding an order-of-magnitude advantage.
In terms of business models, Kimi K3’s launch price is close to Anthropic’s Sonnet series: input pricing is approximately $3 per million tokens, and output pricing is around $15 per million tokens, a roughly threefold increase over the previous Kimi generation. Max Kan argued that if Moonshot AI is not operating at a long-term loss, then Anthropic and OpenAI charging higher prices for models of comparable size suggests their API business may have high profit margins. “Selling API tokens could be more profitable than SaaS,” he said.
However, the two analysts emphasized that these judgments are not based on the AI companies’ public financial data, but rather on reverse inference drawn from Kimi K3’s parameters, pricing, and performance.
7 minutes ago
Goldman Sachs warns that inflationary pressures are spreading across the US, with Fed Chair Walsh facing mounting pressure to raise interest rates.
Goldman Sachs’ latest research report shows that U.S. inflationary pressure is spreading from a narrow set of sectors to a broader range of areas. While current inflation levels have not yet hit their 2022 peak, the expanding scope of price increases is posing greater challenges to the Federal Reserve’s policy efforts. Goldman Sachs economist Jessica Rindels analyzed the extent of inflation spread using the six-month annualized change rate of the Personal Consumption Expenditures (PCE) price index, a key metric closely watched by the Fed. The data shows that, compared to the average inflation level between 1990 and 2019, the pressure index for inflation categories exceeding 3% has reached around 6, while it stood at 10 during the 2022 inflation peak. The report points out that sectors such as audio-visual equipment, financial services, healthcare, and transportation have become key drivers of current price increases. Meanwhile, housing rent inflation, which carries a significant weight in the PCE index, is projected to fall below 3% in the fourth quarter of this year, potentially serving as a key factor easing inflationary pressure. Goldman Sachs’ analysis aligns with recent concerns from new Fed Chair Kevin Warsh about the "broadening of inflation". Warsh stated that preventing price hikes from spreading to more sectors of the economy is a key task for the Federal Reserve. However, unlike former Chair Jerome Powell’s relatively clear policy communication style, Warsh has so far refused to provide specific interest rate path guidance. Jeremy Schwartz, senior U.S. economist at Nomura Securities, noted that the Fed is reducing forward guidance to the market, and this policy uncertainty has heightened concerns on Wall Street. Meanwhile, hawkish voices within the Fed are growing. Dallas Fed President Lorie Logan has expressed support for moderate interest rate hikes, arguing that the current economic resilience is inconsistent with inflation risks.
7 minutes ago
OKX's World Cup prediction campaign has concluded, with the X Layer ecosystem recording over 136 million total transactions in the past 30 days.
According to official data, OKX’s World Cup prediction campaign officially concluded on July 19 alongside the final match, drawing a total of 433,590 participants, with a prize pool of approximately $4.2 million and covering 104 matches. As of July 20, the total number of transactions in OKX’s X Layer ecosystem over the past 30 days reached 136,048,079, with on-chain latency and block settlement frequency both under 100 milliseconds. It is reported that OKX’s World Cup prediction campaign launched on June 3, and was the first self-operated application of Exchange OS, the trading infrastructure built by OKX based on X Layer. Leveraging match prediction as its use case, the campaign further demonstrates X Layer’s capabilities in high-frequency interaction, low-latency settlement, and on-chain trading infrastructure.
7 minutes ago
Hacken: Crypto institutions no longer rely solely on audits; security assessments are shifting to continuous monitoring and operational resilience.
Blockchain security firm Hacken has released a report stating that crypto institutions are re-evaluating project security standards, as traditional smart contract audits and operational histories are no longer sufficient as trust benchmarks. Investors are now focusing on continuous monitoring, signature permission management, and incident response capabilities.
According to Hacken’s Q2 2026 Security & Compliance Report, among the 1,427 projects it tracks, only 9% have deployed third-party continuous monitoring mechanisms, and just 4% have all three elements: monitoring, bug bounty programs, and security audits.
The report shows that of the approximately $764 million in crypto asset losses recorded in Q2 2026, 88.3% came from private key leaks, signature permission issues, and infrastructure security problems, rather than smart contract vulnerabilities.
Hacken noted that institutional investors are increasingly adding assessments of signer changes, collateral support, third-party dependencies, incident response capabilities, and audit scope and timeliness to their due diligence processes.
The report points out that the 14 projects attacked in Q2 had all undergone audits before, but most losses stemmed from areas outside the coverage of traditional audits, including signature devices, cross-chain bridge validation nodes, backend infrastructure, administrator keys, and old contracts still in operation.
Hacken stated that as institutional capital enters the crypto market, projects lacking continuous security proof may face higher risk premiums, fewer investment opportunities, and greater difficulty obtaining insurance and partner support.
Federico Bagiotti, Head of Risk Management at Abraxas Capital, said that whether a project’s security level matches its funding scale has become a key criterion for institutional investment, surpassing project potential.
7 minutes ago
The total accounts on #Tron has surpassed 394M. Active accounts reached 4,646,026, up 4.64% in the past 24 hours.
The total accounts on #Tron has surpassed 394M. Active accounts reached 4,646,026, up 4.64% in the past 24 hours.
7 minutes ago