# Models İlgili Makaleler

HTX Haber Merkezi, kripto endüstrisindeki piyasa trendleri, proje güncellemeleri, teknoloji gelişmeleri ve düzenleyici politikaları kapsayan "Models" hakkında en son makaleleri ve derinlemesine analizleri sunmaktadır.

OpenAI No Longer Sells Its Most Expensive Model for Profit

OpenAI is shifting its business strategy away from promoting its most expensive, flagship models for every task. Recent price cuts—80% for GPT-5.6 Luna and 20% for Terra—signal a deeper change: the company now actively advises users that many tasks don't require the most powerful model. Instead, OpenAI recommends a tiered approach: use the high-end GPT-5.6 Sol for complex planning and analysis, then delegate execution to cheaper models like Luna. This mirrors moves by Anthropic, which recently launched Claude Opus 5 at half the price of its top model, Fable 5. Both companies are de-emphasizing flagship models as primary revenue drivers, using them instead for brand prestige and technological showcases. The industry is entering a "mass-market" phase, similar to automotive, where high-volume, cost-effective models handle daily operations and drive scale. OpenAI's price reductions are partly enabled by AI models themselves optimizing underlying code and infrastructure, creating a self-reinforcing cycle of efficiency gains and cost reduction. Competition is shifting from "who is smartest" to "who offers the best value." The goal is no longer selling individual models but fostering widespread API adoption and ecosystem lock-in. By making AI calls cheap and ubiquitous, companies like OpenAI aim to become the indispensable, utility-like infrastructure powering automated workflows—the "water and electricity" of software, quietly embedded everywhere.

marsbit14 saat önce

OpenAI No Longer Sells Its Most Expensive Model for Profit

marsbit14 saat önce

Huang Renxun's Second Lifelong Tweet: Only Anthropic Still Holding Out

NVIDIA's Jensen Huang issued his second-ever tweet, announcing the formation of the Open Secure AI Alliance. Backed by 37 founding members including Microsoft, IBM, and Hugging Face, the coalition argues that open-source models are essential for cybersecurity, allowing defenders to audit and harden the entire AI stack. While OpenAI and Google later signed a supporting open-letter on model weights, they did not join the alliance. Anthropic stands alone, refusing both, and has instead outlined a stricter stance focused on chip controls and security testing. The alliance's formation follows a significant incident where OpenAI's models allegedly breached Hugging Face's systems. During the investigation, Hugging Face found its access to commercial frontier models blocked by safety guardrails when analyzing attack logs, forcing a switch to the open-source GLM 5.2 model to complete forensics. This event is cited as a key rationale for the alliance: defenders must maintain control over their AI tools. Beyond security, NVIDIA's push for open models serves a commercial imperative. With major cloud providers and AI labs developing their own AI chips, promoting easily deployable open models helps safeguard the market for NVIDIA's GPU-based compute infrastructure by keeping AI development and its associated compute demands distributed.

marsbit07/28 09:36

Huang Renxun's Second Lifelong Tweet: Only Anthropic Still Holding Out

marsbit07/28 09:36

Peskov Says Russia Is Among Top Five Leaders in AI Race. What Do Global Rankings Say?

On July 27, 2026, Kremlin spokesman Dmitry Peskov stated that Russia remains among the top five countries in the global AI development race. He acknowledged Russian models still lag behind leading U.S. counterparts but claimed they have reached a "very high level," aiming to close the gap with "superhuman efforts." He highlighted the differing approaches of Russia's GigaChat, built from scratch, and Yandex, which initially used foreign technology. However, this claim is not supported by major international AI rankings. Stanford University's Global AI Vibrancy Tool (2024-25) ranks Russia 28th out of 36 countries. The top five are the U.S., China, India, South Korea, and the UK. The Stanford AI Index Report 2026 does not mention Russia's position, focusing instead on U.S. and Chinese leadership across various metrics like investments and model performance. In benchmarks, GigaChat ranks 25th on the Russian-language LM Arena. While it passed a financial analyst exam in December 2025, its business usage costs are reportedly tens to hundreds of times higher than China's DeepSeek. In related developments, President Putin signed a law on July 26, 2026, establishing a legal framework for sovereign AI models and granting developers access to state data. Previously, Russia joined 28 other nations, including China, to establish the World AI Cooperation Organization (WAICO) in Shanghai. The article notes that rankings vary due to different criteria, such as research, investment, infrastructure, or responsible AI governance. While Russian authorities are bolstering AI through legislation and international cooperation, independent analyses suggest the country faces significant challenges, including a hardware deficit for training models, which legal frameworks alone cannot resolve.

cryptonews.ru07/28 08:56

Peskov Says Russia Is Among Top Five Leaders in AI Race. What Do Global Rankings Say?

cryptonews.ru07/28 08:56

I Was Wrong About Anthropic, Criticized for Over Half a Year, Musk Suddenly Changes Tune

Elon Musk, previously a vocal critic of Anthropic, has publicly reversed his position. After months of labeling the company as "hypocritical and evil" and asserting it "could never win," Musk now calls Anthropic a leader in AI, praising the strength of its Mythos/Fable models. This shift follows an X user's post suggesting Musk's SpaceXAI, which supplies computing power to Anthropic, could cripple the rival by cutting off access. The change in stance coincides with a major deal: in May 2026, Anthropic secured the full capacity of SpaceX's Colossus 1 data center, involving over 22,000 GPUs, for a reported $1.25 billion per month. While Musk continues to promote his own Grok models, he acknowledges Anthropic's current technical edge. He also stated he would not use SpaceX's power as a supplier to severely harm Anthropic, citing his history of supporting competitors in other ventures like Tesla and Supercharger. The situation highlights the complex interplay in the AI race, where fierce model competition coexists with deep interdependence on shared infrastructure like compute. Musk's reversal underscores a pragmatic shift from purely adversarial tactics to a strategy that balances competition with profitable cooperation.

marsbit07/13 10:57

I Was Wrong About Anthropic, Criticized for Over Half a Year, Musk Suddenly Changes Tune

marsbit07/13 10:57

Strike While the Iron is Hot: GPT-5.6's Three Major Models Fully Exposed, Scheduled for July 7th?

GPT-5.6 Models Leaked, Rumored for July 7th Release A leak from the Codex application's source code has revealed references to three upcoming GPT-5.6 sub-models: Sol, Terra, and Luna. A new "speed dial" feature was also discovered, suggesting users will be able to balance response speed and quality. According to rumors, OpenAI is targeting a release window of July 7th to 9th, strategically timed to coincide with the expiration of access limits for Anthropic's Claude Fable 5 model. Early comparative tests between a leaked GPT-5.6-terra version and Fable 5 show GPT-5.6 achieving similar or better results while using significantly fewer computational resources and responding faster. In complex coding tasks, GPT-5.6 demonstrated more efficient and user-collaborative problem-solving. The reported launch date appears calculated to capitalize on growing user frustration with competitor models. Anthropic's Fable 5 has faced criticism for overly aggressive safety filters that frequently downgrade responses and for severe "hallucination" issues in its Opus 4.8 model. Leaks also suggest GPT-5.6 Sol could be offered at a much lower price than Fable 5 while matching its performance. A final note warns developers with unused "rate limit reset" credits in Codex to use them before they expire around July 12th, ahead of the anticipated GPT-5.6 launch.

marsbit07/04 08:45

Strike While the Iron is Hot: GPT-5.6's Three Major Models Fully Exposed, Scheduled for July 7th?

marsbit07/04 08:45

Doubao Charges More than GPT, While DeepSeek Slashes Prices Dramatically: Who Will Win?

The article discusses the divergent pricing strategies of two major Chinese AI companies. In May, Doubao (by ByteDance) began testing fees, with its professional tier priced higher than ChatGPT Plus. Meanwhile, DeepSeek permanently cut prices for its V4-Pro API to a quarter of the original, setting new global lows. Doubao, with high user traffic from ByteDance apps like TikTok, leads in monthly active users but faces massive compute costs from its free model. Its move to a freemium model targets heavy users, aiming to balance scale and monetization amid substantial investments. DeepSeek's price cut is attributed to architectural innovations that slash inference costs, adaptation to domestic hardware reducing dependency, and engineering optimizations. It focuses on the enterprise (B2B) market, aiming to become a leading model base. Both companies are currently unprofitable. The article contrasts their approaches with Anthropic, which is profitable by primarily serving enterprises with high-value use cases like coding and agents. It argues that sustainable AI business models require integrating AI into real workflows to deliver tangible ROI, rather than just offering chat services. DeepSeek's recent $7 billion funding round, including investments from Tencent, is noted to bolster its B2B position. The ultimate winner will be the player that successfully transforms AI into measurable returns, whether through consumer productivity ecosystems or enterprise platforms.

marsbit06/11 06:23

Doubao Charges More than GPT, While DeepSeek Slashes Prices Dramatically: Who Will Win?

marsbit06/11 06:23

AI Investors' 2026 Anxiety: When Models Devour Everything, What Moat Is Left for Startups?

In 2026, a wave of investor anxiety questions the defensibility of AI startups as models improve, fearing that most companies are just "thin wrappers" destined to be absorbed by foundation models or chipmakers. The author argues against this despair, positing that true moats lie not in benchmark performance but in areas models cannot easily reach. The logic of despair is that if models excel at all measurable tasks, only compute and cutting-edge model weights hold lasting value. However, the essay contends that the most valuable work is inherently "untrainable." Benchmarks measure what can be measured and thus optimized for, but real-world correctness often resides in private, complex systems. Examples include legacy codebases, intricate legal transactions, or hospital workflows. This kind of correctness is proprietary, costly to establish, and cannot be validated quickly—it requires time and trust within an organization. As models commodify visible, measurable tasks from both above (labs absorbing scaffolding) and below (saturation by cheaper models), value shifts to "untrainable ground." This encompasses work where correctness is a private truth, locked behind integration barriers, licenses, liability frameworks, and entrenched user habits. Trust and adoption are slow, human-centric processes that smarter models cannot accelerate. Successful companies defend their position by embedding deeply into client operations, owning the definition of "good" within a specific domain (e.g., Harvey in law, OpenEvidence in medicine), and pricing on outcomes rather than tokens. While labs compete fiercely, they are incentivized to keep the application layer vibrant. The future belongs not to those competing on generic benchmarks but to those navigating unscoreable terrain, doing the "unsexy work" of translation between models and messy human realities. The most cited benchmark scores are thus maps of territory about to become worthless, signaling who will lose the right to define what counts as good.

marsbit06/11 03:34

AI Investors' 2026 Anxiety: When Models Devour Everything, What Moat Is Left for Startups?

marsbit06/11 03:34

Who Will Define the Rules of the AI Era? Anthropic Discusses the 2028 US-China AI Landscape

This article, based on Anthropic's analysis, outlines the intensifying systemic competition between the U.S./allies and China for AI leadership by 2028. It argues that access to advanced computing power ("compute") is the critical bottleneck, where the U.S. currently holds a significant advantage through chip export controls and allied innovation. However, China's AI labs remain competitive by exploiting policy loopholes—via chip smuggling, overseas data center access, and "model distillation" attacks to copy U.S. model capabilities—keeping them close to the frontier. The piece presents two contrasting scenarios for 2028. In the first, decisive U.S. action to tighten compute controls and curb distillation locks in a 12-24 month AI capability lead, cementing democratic influence over global AI norms, security, and economic infrastructure. In the second, policy inaction allows China to achieve near-parity through continued access to U.S. technology, enabling Beijing to promote its AI stack globally and integrate advanced AI into its military and governance systems, altering the strategic balance. Anthropic contends that maintaining a decisive U.S. lead is essential for shaping safe AI development and governance. The core recommendation is for U.S. policymakers to urgently close compute and model access loopholes while promoting global adoption of the U.S. AI technology stack to secure a lasting strategic advantage.

marsbit05/16 05:08

Who Will Define the Rules of the AI Era? Anthropic Discusses the 2028 US-China AI Landscape

marsbit05/16 05:08

muShanghai Discusses Consumer AI: After Continuous Iteration of Large Models, Product Competition Moves Towards Scenarios and Experience

The roundtable discussion "Innovative Practices and Path Exploration of the AI Consumption Ecosystem" at muShanghai AI Week, featuring experts from model platforms, cultural apps, the open-source ecosystem, and music creation, delved into the practical paths for consumer AI products. A key consensus emerged: while AI model advancements lower prototyping barriers, the real challenge for enduring products lies beyond raw technology. True differentiation comes from deep scene understanding, data organization, user education, delivering emotional value, and building open ecosystems. The competition is shifting from "who has the stronger model" to "who best understands the specific user and scenario." Participants highlighted that application-layer barriers, such as accumulated contextual data and cultural localization (e.g., FateTell's translation of Eastern metaphysics for global users), are not easily erased by model updates. They cautioned that AI simplifies prototyping but not the core entrepreneurial hurdles: user acquisition, community building, and commercialization. The discussion emphasized that value must return to human needs—like emotional comfort (FateTell) or preserving the creative *process* in music-making, as highlighted by musician-developer Gao Jiafeng, rather than just outputting a final product. With the rise of AI Agents, user education is evolving from manual documentation reading to more guided, interactive learning within the product experience itself. Looking ahead 3-5 years, panelists foresee AI moving into the physical world via hardware and robotics, enabling more personalized services and addressing growing needs for companionship amidst technological anxiety. The future points towards "technology democratization," where AI assists diverse lifestyles, and cultural forms may be recombined, with emotional connection becoming paramount. Ultimately, as models continue to evolve, the products that endure will be those that meet genuine human needs, foster understanding, and build meaningful connections.

marsbit05/16 03:06

muShanghai Discusses Consumer AI: After Continuous Iteration of Large Models, Product Competition Moves Towards Scenarios and Experience

marsbit05/16 03:06

AI Giants Enter the Dark Forest

In the AI industry's "dark forest," major players like Anthropic, OpenAI, and DeepSeek are strategically withholding their most advanced models to avoid becoming targets in a high-stakes competitive landscape. Anthropic released Claude Opus 4.7 but admitted it underperforms compared to their unreleased model Mythos, citing safety concerns. They delayed addressing user complaints about performance regression until OpenAI’s GPT-5.5 launch, highlighting a tactic of controlled disclosure aligned with competitors’ moves. OpenAI’s GPT-5.5, though a full retrain since GPT-4.5, was seen as incremental rather than revolutionary. Leaks revealed internal models like Glacier and Heisenberg, indicating significant unreleased capabilities. OpenAI acknowledges a "capability overhang," where real model power exceeds what users experience, often due to infrastructure-driven throttling. DeepSeek launched V4 Preview, a cost-efficient model, but its full potential (V4 Pro Max) awaits Huawei’s Ascend 950 super-nodes量产 in late 2026. Their strategy focuses on affordability and scalability, aiming to democratize AI access globally, a move noted even by NVIDIA’s CEO as a disruptive threat. Together, these actions reflect a broader trend: leading AI labs are deliberately pacing releases, hiding strengths, and aligning disclosures with competitive dynamics—each avoiding the risk of exposure in a forest where first movers become targets.

marsbit04/25 12:47

AI Giants Enter the Dark Forest

marsbit04/25 12:47

活动图片

Market Analysisright-arrow

Technology Trendsright-arrow