Mistral has released the weights of Mistral Large 4 on Hugging Face. The model, nicknamed Le Chonk, has approximately 1 trillion parameters with 52 billion activated per step, and supports text and image inputs. On Terminal-Bench 4.0, its score rose from 28% in the preview version to 41%, slightly above GLM-5.3's 40%, but on the new version of the Artificial Analysis Intelligence Index it scored only 26 points, below the 32 points of DeepSeek V4.1 Flash and GLM-5.3. API pricing is $0.70 per million input tokens and $2.10 per million output tokens, under a modified MIT license; using it as a commercial cloud service requires separate authorization.
Mistral has released the weights of Mistral Large 4 on Hugging Face. The model, nicknamed Le Chonk, has approximately 1 trillion parameters with 52 billion activated per step, and supports text and image inputs. On Terminal-Bench 4.0, its score rose from 28% in the preview version to 41%, slightly above GLM-5.3's 40%, but on the new version of the Artificial Analysis Intelligence Index it scored only 26 points, below the 32 points of DeepSeek V4.1 Flash and GLM-5.3. API pricing is $0.70 per mill
The Verge's Decoder podcast invited SemiAnalysis analyst Joey Brookhart to break down the S-1 filed by Anthropic, discussing its potential Nasdaq listing at a valuation of $1.5 trillion to $2 trillion. Brookhart said Anthropic has achieved over $70 billion in ARR, with inference gross margins turning from negative 94% two years ago to positive 85%, and revenue per megawatt of inference at about $65 million, but revenue is highly concentrated in coding scenarios, and about 70% of compute is used for training and research. He also mentioned that Anthropic has $500 billion to $600 billion in compute commitments over the next 7 to 10 years, and compared OpenAI's more consumer-oriented business structure.
The Verge's Decoder podcast invited SemiAnalysis analyst Joey Brookhart to break down the S-1 filed by Anthropic, discussing its potential Nasdaq listing at a valuation of $1.5 trillion to $2 trillion. Brookhart said Anthropic has achieved over $70 billion in ARR, with inference gross margins turning from negative 94% two years ago to positive 85%, and revenue per megawatt of inference at about $65 million, but revenue is highly concentrated in coding scenarios, and about 70% of compute is used
The Google A2UI team and CopilotKit CEO Atai Barkai explain how to build generative UI applications using the A2UI open standard and CopilotKit. A2UI describes interfaces with declarative JSON, and the client only renders components from a trusted component catalog; the same response can be rendered to React, Angular, SwiftUI, Jetpack Compose, and Flutter. The article provides a code example for integrating CopilotKit with Google ADK, and explains the bidirectional flow, implemented via the AG-UI protocol, in which user actions are sent back and the agent regenerates the interface.
The Google A2UI team and CopilotKit CEO Atai Barkai explain how to build generative UI applications using the A2UI open standard and CopilotKit. A2UI describes interfaces with declarative JSON, and the client only renders components from a trusted component catalog; the same response can be rendered to React, Angular, SwiftUI, Jetpack Compose, and Flutter. The article provides a code example for integrating CopilotKit with Google ADK, and explains the bidirectional flow, implemented via the AG-U
Google DeepMind has released Gemini 4 Carbon, which ranks first on the Artificial Analysis Intelligence Index v5 with a score of 59 and leads in 4 of 9 evaluations including HLE-Diamond and TB-Science. The model focuses on cybersecurity and long-horizon tasks, claiming 88.2% on SEC-bench Pro, 54.8% on ExploitGym, and 71.0% on CWE-bench, and is initially available only to trusted cybersecurity personnel through the Fairwind Program. In terms of specifications, it supports a 1M token context and 1M maximum output, with input covering text, images, video, and speech, and output in text; Artificial Analysis lists a standard price of $4/$20 per million input/output tokens, plus a 50% launch discount.
Google DeepMind has released Gemini 4 Carbon, which ranks first on the Artificial Analysis Intelligence Index v5 with a score of 59 and leads in 4 of 9 evaluations including HLE-Diamond and TB-Science. The model focuses on cybersecurity and long-horizon tasks, claiming 88.2% on SEC-bench Pro, 54.8% on ExploitGym, and 71.0% on CWE-bench, and is initially available only to trusted cybersecurity personnel through the Fairwind Program. In terms of specifications, it supports a 1M token context and 1
Cartesia has launched the Managed Agents end-to-end voice agent platform, targeting enterprise call scenarios. On Artificial Analysis's Speech Agent Arena, it ranks first among 30 agents with a preference Elo of 1115. On ServiceNow's EVA-Bench, the platform achieved an accuracy of 87.4% and an experience score of 90.7%, both the highest. The platform co-designs the voice model with the harness, offering capabilities such as 44 languages, tool calling, knowledge bases, and outbound calling batches. The company says it has already supported tens of millions of conversations in finance, retail, hospitality, and healthcare scenarios.
Cartesia has launched the Managed Agents end-to-end voice agent platform, targeting enterprise call scenarios. On Artificial Analysis's Speech Agent Arena, it ranks first among 30 agents with a preference Elo of 1115. On ServiceNow's EVA-Bench, the platform achieved an accuracy of 87.4% and an experience score of 90.7%, both the highest. The platform co-designs the voice model with the harness, offering capabilities such as 44 languages, tool calling, knowledge bases, and outbound calling batche
OpenAI released the GPT-6 Luna (October 29) update, which responds more concisely and reduces overly sycophantic phrasing; under the Preparedness Framework it is still rated High for both cybersecurity and biochemical capabilities, and it uses the same set of safeguards as GPT-6 Luna (October 7). This version improves over the previous version on the gore evaluation in Production Benchmarks, achieving a Pareto improvement between safely handling harmful requests and helpfulness; in the U18 evaluation it improves on sexual content, eating disorders, emotional dependence, and gore, and in the instruction hierarchy evaluation it reaches 99.91% robustness.
OpenAI released the GPT-6 Luna (October 29) update, which responds more concisely and reduces overly sycophantic phrasing; under the Preparedness Framework it is still rated High for both cybersecurity and biochemical capabilities, and it uses the same set of safeguards as GPT-6 Luna (October 7). This version improves over the previous version on the gore evaluation in Production Benchmarks, achieving a Pareto improvement between safely handling harmful requests and helpfulness; in the U18 evalu
Amazon Science and Booking.com combined knowledge distillation with Medusa to compress a 70B-parameter Llama-3.1 teacher model into a 1.1B TinyLlama student model, increasing inference speed by 10 to 20 times and reducing cost to less than one-third while maintaining the teacher model's quality. The approach has been used in three features—AI Trip Planner, Smart Filters, and accommodation description generation—and won the Best Paper Award in the Industry Track at ACL.
Anthropic has upgraded the Claude mobile app. The voice mode on the iPhone version now supports Apple CarPlay, allowing users to ask Claude questions directly while driving. Claude has added 24 new languages across web, desktop, iOS and Android, bringing total support to 33 languages, while voice mode responds faster and recognizes speech more accurately in noisy environments. The iOS and Android versions also support logging in to multiple accounts simultaneously and switching between them directly.
Anthropic has upgraded the Claude mobile app. The voice mode on the iPhone version now supports Apple CarPlay, allowing users to ask Claude questions directly while driving. Claude has added 24 new languages across web, desktop, iOS and Android, bringing total support to 33 languages, while voice mode responds faster and recognizes speech more accurately in noisy environments. The iOS and Android versions also support logging in to multiple accounts simultaneously and switching between them dire
Fliggy Travel AI "Fliggy Bangbang" saw a 118% week-over-week increase in Double 11-related interactions over the past week, and launched a "Double 11 Assistant" that supports cross-product price comparison. The feature has been integrated with all Double 11 promotional products, and one-click room booking has begun gray-scale testing, covering hotel packages and hotel-plus-attraction packages. Users can describe their needs in natural language, and the AI matches products based on context such as membership tier and provides follow-up fulfillment services.
Fliggy Travel AI "Fliggy Bangbang" saw a 118% week-over-week increase in Double 11-related interactions over the past week, and launched a "Double 11 Assistant" that supports cross-product price comparison. The feature has been integrated with all Double 11 promotional products, and one-click room booking has begun gray-scale testing, covering hotel packages and hotel-plus-attraction packages. Users can describe their needs in natural language, and the AI matches products based on context such a
Beijing News Shell Finance reporters tested Alibaba's Qwen Go! and Ant's Lingguang 2.0. The former set up a daily 8 a.m. medication reminder with a single sentence and checked it item by item, while the latter, after authorization on Alipay, sorted out the past six months of bills. The article also reviews the layouts of Meta Muse, OpenAI dots, as well as ByteDance and Tencent, and points out that cross-ecosystem permissions are currently the hardest wall to break down, with ByteDance's SAEP published rejection list reaching 155 apps.
Beijing News Shell Finance reporters tested Alibaba's Qwen Go! and Ant's Lingguang 2.0. The former set up a daily 8 a.m. medication reminder with a single sentence and checked it item by item, while the latter, after authorization on Alipay, sorted out the past six months of bills. The article also reviews the layouts of Meta Muse, OpenAI dots, as well as ByteDance and Tencent, and points out that cross-ecosystem permissions are currently the hardest wall to break down, with ByteDance's SAEP pub
Amazon CEO Andy Jassy said on the Q3 earnings call that AWS third-quarter revenue grew 39.3% year over year to $45.98 billion, and is expected to grow over 40% in both 2027 and 2028, becoming a business with annual revenue exceeding one trillion dollars in the next 5 to 10 years. The company maintained its expectation of about $220 billion in cash capital expenditure for 2026, most of which will go to AWS, and expects spending to increase significantly in 2027. On Bedrock, spending by Fortune 500 companies grew nearly 7x year over year and total tokens grew nearly 17x, revenue commitments for Trainium and Graviton grew more than 20x year over year, and Trainium is planned to be sold externally by the end of 2027.
Amazon CEO Andy Jassy said on the Q3 earnings call that AWS third-quarter revenue grew 39.3% year over year to $45.98 billion, and is expected to grow over 40% in both 2027 and 2028, becoming a business with annual revenue exceeding one trillion dollars in the next 5 to 10 years. The company maintained its expectation of about $220 billion in cash capital expenditure for 2026, most of which will go to AWS, and expects spending to increase significantly in 2027. On Bedrock, spending by Fortune 50
Earnings from Amazon, Intel, and Alphabet successively missed expectations, triggering a collective slump in global tech stocks. Nasdaq 100 index futures extended losses to 2%, and South Korea's KOSPI index fell 4.5% and triggered the SIDECAR mechanism to suspend program selling. Intel plunged 23% after hours, Amazon fell about 7.3% after hours, and Alphabet tumbled 13% in the previous trading session, its largest single-day drop. Market attention is focused on capital expenditures: Amazon's third-quarter spending on equipment and property exceeded $77 billion, and free cash flow over the past 12 months was negative $42.4 billion. CEO Andy Jassy said capital expenditures will increase significantly in 2027.
Earnings from Amazon, Intel, and Alphabet successively missed expectations, triggering a collective slump in global tech stocks. Nasdaq 100 index futures extended losses to 2%, and South Korea's KOSPI index fell 4.5% and triggered the SIDECAR mechanism to suspend program selling. Intel plunged 23% after hours, Amazon fell about 7.3% after hours, and Alphabet tumbled 13% in the previous trading session, its largest single-day drop. Market attention is focused on capital expenditures: Amazon's thi
SanDisk reported revenue of $11 billion for the first quarter of fiscal 2027, up 378% year over year, including $4.8 billion in data center revenue, up 61% quarter over quarter. Management said customers have underestimated the NAND capacity needed to run AI infrastructure; about 50% of 2027 supply and roughly two-thirds of 2028 supply have already been booked under NBM long-term orders, with supply visibility extending to 2030 and even further out. The first HBF memory chip has completed tape-out, with the first samples planned for delivery to customers next year.
SanDisk reported revenue of $11 billion for the first quarter of fiscal 2027, up 378% year over year, including $4.8 billion in data center revenue, up 61% quarter over quarter. Management said customers have underestimated the NAND capacity needed to run AI infrastructure; about 50% of 2027 supply and roughly two-thirds of 2028 supply have already been booked under NBM long-term orders, with supply visibility extending to 2030 and even further out. The first HBF memory chip has completed tape-o
Zhidemai Technology's AI-related business revenue in the first three quarters of 2026 reached 108 million yuan, an increase of about 233% from 32.2905 million yuan in the same period last year, with its share of total revenue rising from 4% to 13%. The company's revenue in the same period was 828 million yuan, up 2.79% year-on-year, with a net loss attributable to shareholders of 52.5058 million yuan. Its AI services have been deployed via the "What's Worth Buying Shopping Manager" Skill on Tencent WorkBuddy and HarmonyOS Xiaoyi Claw; the "Haina" MCP Server delivered 3.39 billion outputs in the first three quarters, and the company launched the brand analysis product "Zhishu Matrix" and the enterprise model service platform OpenHubs.
Zhidemai Technology's AI-related business revenue in the first three quarters of 2026 reached 108 million yuan, an increase of about 233% from 32.2905 million yuan in the same period last year, with its share of total revenue rising from 4% to 13%. The company's revenue in the same period was 828 million yuan, up 2.79% year-on-year, with a net loss attributable to shareholders of 52.5058 million yuan. Its AI services have been deployed via the "What's Worth Buying Shopping Manager" Skill on Tenc
Canadian AI model developer Cohere is pursuing a funding round of up to $5 billion, with a pre-money valuation of $20.5 billion, roughly three times that of a year ago. The company said it has completed the first closing, with investors including the Canadian government and Nvidia making funding commitments, among which the Canadian government agreed to invest up to C$1.6 billion (about $1.14 billion). Cohere plans to keep its headquarters in Canada and expects to invest about $1.2 billion in the country by 2030 to build new computing infrastructure.
Canadian AI model developer Cohere is pursuing a funding round of up to $5 billion, with a pre-money valuation of $20.5 billion, roughly three times that of a year ago. The company said it has completed the first closing, with investors including the Canadian government and Nvidia making funding commitments, among which the Canadian government agreed to invest up to C$1.6 billion (about $1.14 billion). Cohere plans to keep its headquarters in Canada and expects to invest about $1.2 billion in th
Amazon's Q3 2026 earnings show AWS revenue of $46 billion, up 39.3% year over year, with $578 billion in backlog orders, and AI inference revenue accounting for more than half of AWS's AI revenue. Management expects AWS to grow at an annual rate above 40% in 2027-2028 and to approach $1 trillion in revenue within 5 to 10 years, but capital expenditures will be about $220 billion in 2026 and will rise significantly further in 2027, sending the company's shares down 7.3% after hours. The earnings call also mentioned that spending by Fortune 500 companies on Bedrock grew nearly 7x year over year, token call volume grew nearly 17x, and Trainium is planned to be sold externally by the end of 2027.
Amazon's Q3 2026 earnings show AWS revenue of $46 billion, up 39.3% year over year, with $578 billion in backlog orders, and AI inference revenue accounting for more than half of AWS's AI revenue. Management expects AWS to grow at an annual rate above 40% in 2027-2028 and to approach $1 trillion in revenue within 5 to 10 years, but capital expenditures will be about $220 billion in 2026 and will rise significantly further in 2027, sending the company's shares down 7.3% after hours. The earnings
Google will gradually shut down Jules, the AI coding agent launched in December 2024, and fold the lessons accumulated from the tool into its AI coding platform Antigravity. According to Jules' official X account, Jules began as an experiment exploring how asynchronous coding agents could help developers ship faster, and its lessons have already influenced multiple Google products. The integrated direction is to let users both autonomously execute background tasks in remote environments and use an agent-centric development environment.
Google will gradually shut down Jules, the AI coding agent launched in December 2024, and fold the lessons accumulated from the tool into its AI coding platform Antigravity. According to Jules' official X account, Jules began as an experiment exploring how asynchronous coding agents could help developers ship faster, and its lessons have already influenced multiple Google products. The integrated direction is to let users both autonomously execute background tasks in remote environments and use
Michael Pfaffenberger of Pydantic used 5 small Pydantic AI applications and 8 models to conduct 5,276 scored runs, comparing code mode with direct tool calling. On long tasks, code mode reduced average cost by 58%, latency by 51%, and increased success rate by 16 percentage points; however, for short tasks with 5 items, 3 out of 5 tasks were instead 21% to 38% more expensive. The author argues that models are better at writing loops than at acting as the loop, and the gap widens as tasks get longer.
Michael Pfaffenberger of Pydantic used 5 small Pydantic AI applications and 8 models to conduct 5,276 scored runs, comparing code mode with direct tool calling. On long tasks, code mode reduced average cost by 58%, latency by 51%, and increased success rate by 16 percentage points; however, for short tasks with 5 items, 3 out of 5 tasks were instead 21% to 38% more expensive. The author argues that models are better at writing loops than at acting as the loop, and the gap widens as tasks get lon
Reports from Grist and WIRED point out that although most emergency management agencies do not allow employees to use generative AI, experts believe that state and local disaster departments are already using it privately to draft alerts. A county emergency manager in Alabama once used AI to generate a hurricane risk map, but the result was meteorologically inaccurate; meanwhile, models from companies such as EM1 can already rewrite flood alerts for the San Juan River in New Mexico, adding the issuing authority, location, and evacuation guidance. Experts warn that inexperienced emergency personnel find it difficult to judge model output, and public distrust of AI and government may also backfire on the credibility of alerts.
Reports from Grist and WIRED point out that although most emergency management agencies do not allow employees to use generative AI, experts believe that state and local disaster departments are already using it privately to draft alerts. A county emergency manager in Alabama once used AI to generate a hurricane risk map, but the result was meteorologically inaccurate; meanwhile, models from companies such as EM1 can already rewrite flood alerts for the San Juan River in New Mexico, adding the i
Anthropic hiring manager Felix Reiseberg announced that the company will no longer use traditional coding interviews to evaluate candidates, citing the fact that Claude is already stronger than humans at both React challenges and assembly problems, and that truly impactful builders do not win through such tests. He advises junior engineers to pick an idea that can drive business impact, build a working prototype over the weekend, focus on just one good idea, and leverage their fresh perspective on users. On the same day, Google DeepMind released the Genie 4 world model, and Claude Opus 5.5 (58 points) topped the Artificial Analysis Intelligence Index v5.
Anthropic hiring manager Felix Reiseberg announced that the company will no longer use traditional coding interviews to evaluate candidates, citing the fact that Claude is already stronger than humans at both React challenges and assembly problems, and that truly impactful builders do not win through such tests. He advises junior engineers to pick an idea that can drive business impact, build a working prototype over the weekend, focus on just one good idea, and leverage their fresh perspective
