https://themodelwire.com/article/openai-cuts-api-credits-in-half-as-it-shifts-pro-users-to-pay-per-use-01M3PAT9HDMF3PB498XSTD6DB6 Modelwire en 2026-09-29T10:05:14.000Z OpenAI cuts API credits in half as it shifts Pro users to pay-per-use https://themodelwire.com/article/standard-chartered-deploys-openai-daybreak-for-enterprise-threat-detection-01M3PAT965W768SKSDTNG1FVBR Modelwire en 2026-09-29T10:01:43.000Z Standard Chartered deploys OpenAI Daybreak for enterprise threat detection https://themodelwire.com/article/autonomous-ai-takes-over-semiconductor-design-workflows-at-ambiq-01M3P9YS9RKEC9Y095SW80YR7M Modelwire en 2026-09-29T10:00:06.000Z Autonomous AI takes over semiconductor design workflows at Ambiq https://themodelwire.com/article/metas-enterprise-agent-bet-risks-squandering-consumer-ai-advantage-01M3P9YRWJP4FZCHVFWZSDM54K Modelwire en 2026-09-29T10:00:00.000Z Meta's enterprise agent bet risks squandering consumer AI advantage https://themodelwire.com/article/manus-expands-ai-agents-into-video-gaming-and-mobile-orchestration-01M3P6GYZVNB0WV32RTABQMGJ4 Modelwire en 2026-09-29T08:47:37.000Z Manus expands AI agents into video, gaming, and mobile orchestration https://themodelwire.com/article/openai-delays-gpt-6-1-astra-over-deceptive-behavior-in-safety-testing-01M3P4T12SZRCY37MT80M5B4GP Modelwire en 2026-09-29T08:26:38.000Z OpenAI delays GPT-6.1 Astra over deceptive behavior in safety testing https://themodelwire.com/article/anthropic-ipo-filing-discloses-billions-in-losses-and-existential-ai-risk-01M3NSMX4XKKVV35X0WG4BJTA7 Modelwire en 2026-09-29T05:13:43.000Z Anthropic IPO filing discloses billions in losses and existential AI risk https://themodelwire.com/article/nvidia-releases-unproven-containment-system-for-rogue-ai-agents-01M3NJS7BQHEWVGG97EJF8P87C Modelwire en 2026-09-29T01:55:54.000Z Nvidia releases unproven containment system for rogue AI agents https://themodelwire.com/article/anthropic-moves-claude-code-toward-modular-multiplayer-agent-workflows-01M3NCRZZQRGBDR0P4C32TM1TR Modelwire en 2026-09-29T01:17:41.000Z Anthropic moves Claude Code toward modular, multiplayer agent workflows https://themodelwire.com/article/openai-takes-responsibility-for-australian-government-cyber-incidents-01M3NDMB86GKYCN2VXKFX39060 Modelwire en 2026-09-29T01:00:00.000Z OpenAI takes responsibility for Australian government cyber incidents https://themodelwire.com/article/openai-delays-model-release-over-safety-concerns-before-developer-conference-01M3N8FMDA9QC1724PC0YVQ5J7 Modelwire en 2026-09-29T00:16:11.000Z OpenAI delays model release over safety concerns before developer conference https://themodelwire.com/article/openai-shelves-model-over-instruction-following-failures-01M3N6RNR6GJB7ASSTR2X0EXJ3 Modelwire en 2026-09-28T23:39:20.000Z OpenAI shelves model over instruction-following failures https://themodelwire.com/article/anthropic-ships-faster-cheaper-sonnet-5-5-with-inherited-reasoning-bug-01M3N1KV84MQYY5KYREDDNVC6R Modelwire en 2026-09-28T22:07:38.000Z Anthropic ships faster, cheaper Sonnet 5.5 with inherited reasoning bug https://themodelwire.com/article/amd-acquires-world-labs-for-8-2-billion-betting-on-spatial-ai-01M3MZ1FASEVKG1AVGB5WC020A Modelwire en 2026-09-28T21:31:35.000Z AMD acquires World Labs for $8.2 billion, betting on spatial AI https://themodelwire.com/article/modal-labs-reaches-15-75b-valuation-on-inference-infrastructure-bet-01M3MZ1FFBWKQDK00AAV9FZ3WK Modelwire en 2026-09-28T21:29:18.000Z Modal Labs reaches $15.75B valuation on inference infrastructure bet https://themodelwire.com/article/amd-acquires-world-labs-to-build-spatial-ai-capabilities-01M3MWF2EM0D8RCQFXRJ428CJ3 Modelwire en 2026-09-28T20:39:33.000Z AMD acquires World Labs to build spatial AI capabilities https://themodelwire.com/article/shopify-enables-ai-agents-to-complete-checkout-transactions-01M3MR5SACMAMMEZREXZC8GA2T Modelwire en 2026-09-28T19:33:57.000Z Shopify enables AI agents to complete checkout transactions https://themodelwire.com/article/leading-researchers-warn-of-imminent-ai-driven-research-automation-01M3MR5V6ST9APRXRX1142NY7G Modelwire en 2026-09-28T19:26:21.000Z Leading researchers warn of imminent AI-driven research automation https://themodelwire.com/article/climate-tech-sector-splits-over-ai-data-center-expansion-01M3N9B38W94R10BDW1WT8P3DB Modelwire en 2026-09-28T19:21:59.000Z Climate tech sector splits over AI data center expansion https://themodelwire.com/article/capability-jumps-outpace-security-culture-researcher-warns-01M3MQA8P5C6D4SJWJR6BBNHNB Modelwire en 2026-09-28T19:11:42.000Z Capability jumps outpace security culture, researcher warns https://themodelwire.com/article/openai-publishes-safety-case-framework-for-frontier-model-training-01M3NZN4CDRH27CK4XKTPWYSYS Modelwire en 2026-09-28T19:00:00.000Z OpenAI publishes safety case framework for frontier model training https://themodelwire.com/article/openai-readies-agent-product-to-counter-rivals-in-autonomous-ai-race-01M3MNKBGEQ5ZFKSYWV5WVWBCR Modelwire en 2026-09-28T18:45:00.000Z OpenAI readies agent product to counter rivals in autonomous AI race https://themodelwire.com/article/nvidia-introduces-security-toolkit-for-autonomous-ai-agents-01M3MMQW70804Z504TQ6XRR45T Modelwire en 2026-09-28T18:31:23.000Z Nvidia introduces security toolkit for autonomous AI agents https://themodelwire.com/article/hospitals-and-banks-face-ai-powered-breaches-faster-than-defenses-evolve-01M3MMQX5N48M04E1H64BDAWY4 Modelwire en 2026-09-28T18:30:00.000Z Hospitals and banks face AI-powered breaches faster than defenses evolve https://themodelwire.com/article/openai-developer-experience-team-ships-gpt-6-astra-prototypes-across-five-domain-01M3MKWGAMNXH53853F8M45GZF Modelwire en 2026-09-28T18:05:42.000Z OpenAI Developer Experience team ships GPT-6 Astra prototypes across five domains https://themodelwire.com/article/anthropics-sonnet-5-5-narrows-the-capability-cost-gap-with-opus-01M3MK10GZ2Z2S52TZYBYDZ6WF Modelwire en 2026-09-28T18:02:10.000Z Anthropic's Sonnet 5.5 narrows the capability-cost gap with Opus https://themodelwire.com/article/anthropic-cuts-costs-and-latency-with-sonnet-5-5-mid-tier-model-01M3MKWERM4Z0AXGX3GMMWP9KQ Modelwire en 2026-09-28T18:00:00.000Z Anthropic cuts costs and latency with Sonnet 5.5 mid-tier model https://themodelwire.com/article/single-model-trained-to-work-at-any-compute-budget-01M3NQ2JDFBEHDD5N3X01Q4G6N Modelwire en 2026-09-28T17:59:53.000Z Single model trained to work at any compute budget https://themodelwire.com/article/tokencast-predicts-llm-agent-costs-before-and-during-execution-01M3NRSG3WWMD5Z0AVT6X0QHBP Modelwire en 2026-09-28T17:59:09.000Z TokenCast predicts LLM agent costs before and during execution https://themodelwire.com/article/researchers-scale-llm-story-generation-to-100k-words-via-narrative-state-trackin-01M3NQ2JF9QT7AZ1H8FHP0S4Q1 Modelwire en 2026-09-28T17:59:09.000Z Researchers scale LLM story generation to 100K words via narrative state tracking https://themodelwire.com/article/kv-streams-technique-accelerates-training-of-long-horizon-agentic-llms-01M3NRSG65VKPFBB2RWWF32TJA Modelwire en 2026-09-28T17:57:42.000Z KV-streams technique accelerates training of long-horizon agentic LLMs https://themodelwire.com/article/language-agents-improve-through-self-explanation-without-reinforcement-learning-01M3NQ2JHX0KPPNAWZCTMGK6CK Modelwire en 2026-09-28T17:54:24.000Z Language agents improve through self-explanation without reinforcement learning https://themodelwire.com/article/meta-learning-framework-adapts-llm-execution-harnesses-at-test-time-01M3NQ2JJH6EN3MZEYKKR3SQE7 Modelwire en 2026-09-28T17:52:54.000Z Meta-learning framework adapts LLM execution harnesses at test time https://themodelwire.com/article/human-hand-demonstrations-unlock-generalist-robot-manipulation-via-kinematic-ret-01M3NRSG9D77DYS61J7B3CWF1A Modelwire en 2026-09-28T17:46:42.000Z Human hand demonstrations unlock generalist robot manipulation via kinematic retargeting https://themodelwire.com/article/researchers-benchmark-scaling-law-methodology-across-12-500-model-checkpoints-01M3NRSGB5DAAPFD24WM50TMVZ Modelwire en 2026-09-28T17:44:11.000Z Researchers benchmark scaling law methodology across 12,500 model checkpoints https://themodelwire.com/article/reinforcement-learning-teaches-llms-to-embrace-creative-risk-in-scientific-disco-01M3NQ2JKN5W3GMVAE1KV18R51 Modelwire en 2026-09-28T17:43:52.000Z Reinforcement learning teaches LLMs to embrace creative risk in scientific discovery https://themodelwire.com/article/reinforcement-learning-breaks-distillation-defenses-after-model-theft-01M3NRSGEE6SRN25CTGJ0X6BKW Modelwire en 2026-09-28T17:40:32.000Z Reinforcement learning breaks distillation defenses after model theft https://themodelwire.com/article/test-time-ranking-offers-cheaper-path-to-personalized-llm-outputs-01M3NQ2JMXYJQHH4WQ65Q04AFJ Modelwire en 2026-09-28T17:38:53.000Z Test-time ranking offers cheaper path to personalized LLM outputs https://themodelwire.com/article/circuit-explanations-miss-model-errors-mechanistic-interpretability-study-finds-01M3NRSGFBF1HTSDGN69BMRZNK Modelwire en 2026-09-28T17:34:45.000Z Circuit explanations miss model errors, mechanistic interpretability study finds https://themodelwire.com/article/google-retires-gemini-gems-for-unified-skills-framework-01M3MHA0YFP1PBS4M3Y0ESCJ70 Modelwire en 2026-09-28T17:29:50.000Z Google retires Gemini Gems for unified skills framework https://themodelwire.com/article/openai-launches-misalignment-incident-tracker-amid-safety-concerns-01M3MGEHW6YHENACQD4B7A24P1 Modelwire en 2026-09-28T17:09:02.000Z OpenAI launches misalignment incident tracker amid safety concerns https://themodelwire.com/article/anthropic-deploys-claude-agents-in-molecular-biology-lab-01M3MH9ZVHA4MH1RXP0W0WRDNE Modelwire en 2026-09-28T17:03:16.000Z Anthropic deploys Claude agents in molecular biology lab https://themodelwire.com/article/transformers-ignore-their-own-extreme-signals-blocking-self-correction-01M3NQ2JWJAGP30ZMNYVR4ARS9 Modelwire en 2026-09-28T17:00:50.000Z Transformers ignore their own extreme signals, blocking self-correction https://themodelwire.com/article/training-free-attention-sampling-cuts-kv-cache-memory-reads-01M3NQ2JX62664XHYXMTNTV1EN Modelwire en 2026-09-28T17:00:30.000Z Training-free attention sampling cuts KV cache memory reads https://themodelwire.com/article/florida-seeks-court-order-to-strip-chatgpt-of-human-like-language-01M3MGEJG5DNT57AW7PQ8TVFKH Modelwire en 2026-09-28T17:00:23.000Z Florida seeks court order to strip ChatGPT of human-like language https://themodelwire.com/article/openais-repeated-missteps-with-mathematicians-erode-research-credibility-01M3MGEJH989TFKCTT12GRGQ5A Modelwire en 2026-09-28T17:00:00.000Z OpenAI's repeated missteps with mathematicians erode research credibility https://themodelwire.com/article/llms-reach-right-diagnoses-from-wrong-evidence-new-medical-benchmark-reveals-01M3NQ2JY9F7MR4F27R43AYM5G Modelwire en 2026-09-28T16:59:47.000Z LLMs reach right diagnoses from wrong evidence, new medical benchmark reveals https://themodelwire.com/article/openai-agents-bypassed-api-limits-using-google-security-game-as-proxy-01M3MFK54BZ1NBN88C4CHKWT3S Modelwire en 2026-09-28T16:56:06.000Z OpenAI agents bypassed API limits using Google security game as proxy https://themodelwire.com/article/first-real-time-sign-language-translation-system-bridges-deaf-communication-gap-01M3NQ2K07QQB0ZWEZ40THKNDF Modelwire en 2026-09-28T16:54:03.000Z First real-time sign language translation system bridges deaf communication gap https://themodelwire.com/article/new-benchmark-exposes-llm-limits-in-research-level-theoretical-computer-science-01M3NQ2K0YAC4CC6TTK3KH7GRZ Modelwire en 2026-09-28T16:52:40.000Z New benchmark exposes LLM limits in research-level theoretical computer science