# \#data-points **URL:** https://community.deeplearning.ai/tag/data-points/54.md [Latest](https://community.deeplearning.ai/latest.md) · [Categories](https://community.deeplearning.ai/categories.md) · [Tags](https://community.deeplearning.ai/tags.md) --- ## [Should coding-agent benchmarks report task shape, not only aggregate Pass@1?](https://community.deeplearning.ai/t/should-coding-agent-benchmarks-report-task-shape-not-only-aggregate-pass-1/893900)
**Author:** [@Yu\_LIU5](https://community.deeplearning.ai/u/Yu_LIU5)\ **Replies:** 0\ **Last updated:** [July 19, 2026, 3:02pm UTC](https://community.deeplearning.ai/t/should-coding-agent-benchmarks-report-task-shape-not-only-aggregate-pass-1/893900 "2026-07-19T15:02:28Z")
Disclosure: I maintain Tura. This is a disclosed technical analysis of public benchmark artifacts, not an independent product review. A question about how we evaluate high-effort coding-agent modes: is one aggregate Pas… --- ## [What is poaching?](https://community.deeplearning.ai/t/what-is-poaching/892910)
**Author:** [@abitrolly](https://community.deeplearning.ai/u/abitrolly)\ **Replies:** 2\ **Last updated:** [June 18, 2026, 9:16am UTC](https://community.deeplearning.ai/t/what-is-poaching/892910 "2026-06-18T09:16:08Z")
In Direct Solicitation and Links to External Platforms there is a reference that poaching is prohibited. But poaching is killing rare animals for selling. I don’t understand. The guidelines could explain this predator te… --- ## [OpenAI launches apps inside ChatGPT, with new Apps SDK](https://community.deeplearning.ai/t/openai-launches-apps-inside-chatgpt-with-new-apps-sdk/880886)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [October 6, 2025, 8:58pm UTC](https://community.deeplearning.ai/t/openai-launches-apps-inside-chatgpt-with-new-apps-sdk/880886 "2025-10-06T20:58:23Z")
Subscribe for free access to :arrow\_forward: Data Points! At its DevDay, OpenAI demonstrated a new feature that lets users work with third-party applications directly within ChatGPT conversations. Users can tag apps … --- ## [DeepSeek unveils sparse attention model for cheaper long-context inference](https://community.deeplearning.ai/t/deepseek-unveils-sparse-attention-model-for-cheaper-long-context-inference/880690)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [October 3, 2025, 6:40pm UTC](https://community.deeplearning.ai/t/deepseek-unveils-sparse-attention-model-for-cheaper-long-context-inference/880690 "2025-10-03T18:40:00Z")
Subscribe for free access to :arrow\_forward: Data Points! DeepSeek released V3.2-exp, an experimental model with a new sparse attention system that cuts inference costs for long-context operations by up to 50 percent… --- ## [Anthropic’s Claude Sonnet 4.5 brings major upgrades for coding](https://community.deeplearning.ai/t/anthropic-s-claude-sonnet-4-5-brings-major-upgrades-for-coding/880488)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [September 30, 2025, 2:30pm UTC](https://community.deeplearning.ai/t/anthropic-s-claude-sonnet-4-5-brings-major-upgrades-for-coding/880488 "2025-09-30T14:30:13Z")
Subscribe for free access to :arrow\_forward: Data Points! Claude Sonnet 4.5 now leads on coding benchmarks like SWE-bench Verified and OSWorld, reflecting improved software engineering and computer task performance. … --- ## [Nvidia and OpenAI plan $100 billion AI infrastructure partnership](https://community.deeplearning.ai/t/nvidia-and-openai-plan-100-billion-ai-infrastructure-partnership/879912)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 1\ **Last updated:** [September 29, 2025, 10:46pm UTC](https://community.deeplearning.ai/t/nvidia-and-openai-plan-100-billion-ai-infrastructure-partnership/879912 "2025-09-29T22:46:48Z")
Subscribe for free access to :arrow\_forward: Data Points! The two AI giants announced plans for a strategic partnership to deploy at least 10 gigawatts of Nvidia systems for OpenAI’s next-generation AI infrastructure… --- ## [OpenAI releases GPT-5-Codex with platform updates for developers](https://community.deeplearning.ai/t/openai-releases-gpt-5-codex-with-platform-updates-for-developers/879715)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 1\ **Last updated:** [September 23, 2025, 5:04am UTC](https://community.deeplearning.ai/t/openai-releases-gpt-5-codex-with-platform-updates-for-developers/879715 "2025-09-23T05:04:57Z")
Subscribe for free access to :arrow\_forward: Data Points! GPT-5-Codex, a specialized coding version of GPT-5, outperforms the base model on code refactoring tasks (51.3 percent vs. 33.9 percent) and can work independ… --- ## [Qwen3-Next employs hybrid attention for long context inputs](https://community.deeplearning.ai/t/qwen3-next-employs-hybrid-attention-for-long-context-inputs/879461)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [September 16, 2025, 1:10pm UTC](https://community.deeplearning.ai/t/qwen3-next-employs-hybrid-attention-for-long-context-inputs/879461 "2025-09-16T13:10:39Z")
Subscribe for free access to :arrow\_forward: Data Points! Alibaba introduced Qwen3-Next-80B-A3B, which activates only 3 billion of its 80 billion parameters during inference using a sparse mixture-of-experts design. … --- ## [✨ New course! Enroll in Knowledge Graphs for AI Agent API Discovery](https://community.deeplearning.ai/t/new-course-enroll-in-knowledge-graphs-for-ai-agent-api-discovery/879040)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [September 10, 2025, 5:54pm UTC](https://community.deeplearning.ai/t/new-course-enroll-in-knowledge-graphs-for-ai-agent-api-discovery/879040 "2025-09-10T17:54:55Z")
:arrow\_forward: Enroll Now! Learn how to help AI agents find and execute the right APIs in the right order using a knowledge graph in “Knowledge Graphs for AI Agent API Discovery”, taught by Pavithra G K (Head of Bus… --- ## [Anthropic and Authors’ Guild settle copyright lawsuit](https://community.deeplearning.ai/t/anthropic-and-authors-guild-settle-copyright-lawsuit/878874)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [September 8, 2025, 9:09pm UTC](https://community.deeplearning.ai/t/anthropic-and-authors-guild-settle-copyright-lawsuit/878874 "2025-09-08T21:09:38Z")
Subscribe for free access to :arrow\_forward: Data Points! Anthropic agreed to pay roughly $1.5 billion to settle a copyright infringement lawsuit brought by authors, compensating $3,000 per book for an estimated 500,000… --- ## [Swiss universities unveil fully open source AI model](https://community.deeplearning.ai/t/swiss-universities-unveil-fully-open-source-ai-model/878662)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [September 6, 2025, 4:44pm UTC](https://community.deeplearning.ai/t/swiss-universities-unveil-fully-open-source-ai-model/878662 "2025-09-06T16:44:07Z")
Subscribe for free access to :arrow\_forward: Data Points! Switzerland launched Apertus, a national LLM developed by EPFL, ETH Zurich, and the Swiss National Supercomputing Centre as an alternative to ChatGPT, Meta’s … --- ## [OpenAI releases gpt-realtime and updates API for voice applications](https://community.deeplearning.ai/t/openai-releases-gpt-realtime-and-updates-api-for-voice-applications/878021)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [August 29, 2025, 9:08pm UTC](https://community.deeplearning.ai/t/openai-releases-gpt-realtime-and-updates-api-for-voice-applications/878021 "2025-08-29T21:08:27Z")
Subscribe for free access to :arrow\_forward: Data Points! OpenAI launched “gpt-realtime,” a new speech-to-speech model that processes audio directly through a single model rather than chaining multiple models togethe… --- ## [Perplexity’s new subscription pays publishers for AI traffic, not just clicks](https://community.deeplearning.ai/t/perplexity-s-new-subscription-pays-publishers-for-ai-traffic-not-just-clicks/877722)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [August 25, 2025, 6:11pm UTC](https://community.deeplearning.ai/t/perplexity-s-new-subscription-pays-publishers-for-ai-traffic-not-just-clicks/877722 "2025-08-25T18:11:24Z")
Subscribe for free access to :arrow\_forward: Data Points! Perplexity announced Comet Plus, a $5 monthly subscription service that offers access to premium content from trusted publishers and journalists while introdu… --- ## [New Arabic language model outperforms larger competitors](https://community.deeplearning.ai/t/new-arabic-language-model-outperforms-larger-competitors/821081)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 2\ **Last updated:** [August 21, 2025, 3:40pm UTC](https://community.deeplearning.ai/t/new-arabic-language-model-outperforms-larger-competitors/821081 "2025-08-21T15:40:57Z")
Subscribe for free access to :arrow\_forward: Data Points! The Technology Innovation Institute released Falcon-Arabic, a 7B parameter language model built on the Falcon 3 architecture. The model handles Arabic, English, … --- ## [MolmoAct introduces action reasoning models for more explainable robot control](https://community.deeplearning.ai/t/molmoact-introduces-action-reasoning-models-for-more-explainable-robot-control/873827)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [August 15, 2025, 7:53pm UTC](https://community.deeplearning.ai/t/molmoact-introduces-action-reasoning-models-for-more-explainable-robot-control/873827 "2025-08-15T19:53:34Z")
Subscribe for free access to :arrow\_forward: Data Points! Researchers from the University of Washington and Allen Institute for AI have developed MolmoAct, a family of open-source robotic foundation models that integ… --- ## [AI-generated rock band covertly gains 500,000 Spotify listeners](https://community.deeplearning.ai/t/ai-generated-rock-band-covertly-gains-500-000-spotify-listeners/847283)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [July 7, 2025, 8:19pm UTC](https://community.deeplearning.ai/t/ai-generated-rock-band-covertly-gains-500-000-spotify-listeners/847283 "2025-07-07T20:19:20Z")
Subscribe for free access to :arrow\_forward: Data Points! A fake band called Velvet Sundown drew more than 500,000 Spotify listeners within weeks of releasing two rock-style albums. Online sleuths found no record of the… --- ## [Google previews upgraded Gemini 2.5 Pro, spotlighting coding gains](https://community.deeplearning.ai/t/google-previews-upgraded-gemini-2-5-pro-spotlighting-coding-gains/828803)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [June 6, 2025, 7:45pm UTC](https://community.deeplearning.ai/t/google-previews-upgraded-gemini-2-5-pro-spotlighting-coding-gains/828803 "2025-06-06T19:45:24Z")
Subscribe for free access to :arrow\_forward: Data Points! Google introduced an enhanced preview of Gemini 2.5 Pro, which achieved a 24-point Elo score improvement on LMArena (reaching 1470) and a 35-point jump on WebDev… --- ## [Perplexity’s Labs lets users create reports, apps, and dashboards](https://community.deeplearning.ai/t/perplexity-s-labs-lets-users-create-reports-apps-and-dashboards/826312)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [June 2, 2025, 9:01pm UTC](https://community.deeplearning.ai/t/perplexity-s-labs-lets-users-create-reports-apps-and-dashboards/826312 "2025-06-02T21:01:56Z")
Subscribe for free access to :arrow\_forward: Data Points! Perplexity introduced Labs, a new feature that enables Pro subscribers to use AI-based research and analysis to generate complete projects including reports, spr… --- ## [Microsoft’s Aurora AI model outperforms numerical Earth system forecasts](https://community.deeplearning.ai/t/microsoft-s-aurora-ai-model-outperforms-numerical-earth-system-forecasts/822799)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 1\ **Last updated:** [May 28, 2025, 4:32am UTC](https://community.deeplearning.ai/t/microsoft-s-aurora-ai-model-outperforms-numerical-earth-system-forecasts/822799 "2025-05-28T04:32:31Z")
Subscribe for free access to :arrow\_forward: Data Points! Microsoft Research introduced Aurora, a versatile model trained on over one million hours of diverse geophysical data. Researchers claim Aurora can predict weath… --- ## [Meta releases chemistry research data set and model](https://community.deeplearning.ai/t/meta-releases-chemistry-research-data-set-and-model/818077)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [May 19, 2025, 6:05pm UTC](https://community.deeplearning.ai/t/meta-releases-chemistry-research-data-set-and-model/818077 "2025-05-19T18:05:11Z")
Subscribe for free access to :arrow\_forward: Data Points! Meta released a new data set called Open Molecules 2025 (OMol25), created through 6 billion compute hours and 100 million quantum mechanical calculations. The co… --- ## [U.S. Copyright Office releases AI fair use report amid leadership upheaval](https://community.deeplearning.ai/t/u-s-copyright-office-releases-ai-fair-use-report-amid-leadership-upheaval/814232)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [May 12, 2025, 9:35pm UTC](https://community.deeplearning.ai/t/u-s-copyright-office-releases-ai-fair-use-report-amid-leadership-upheaval/814232 "2025-05-12T21:35:53Z")
Subscribe for free access to :arrow\_forward: Data Points! The U.S. Copyright Office quietly posted a pre-publication version of its AI and fair use report just one day before Register of Copyrights Shira Perlmutter was … --- ## [Google Labs uses Gemini to create Little Language Lessons](https://community.deeplearning.ai/t/google-labs-uses-gemini-to-create-little-language-lessons/810754)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [May 5, 2025, 9:07pm UTC](https://community.deeplearning.ai/t/google-labs-uses-gemini-to-create-little-language-lessons/810754 "2025-05-05T21:07:29Z")
What are your reactions? :megaphone: Comment! Google engineers developed three experimental language learning tools powered by the Gemini API. The “Little Language Lessons” collection includes Tiny Lesson, which provide… --- ## [OpenAI rolls back sycophantic GPT-4o update](https://community.deeplearning.ai/t/openai-rolls-back-sycophantic-gpt-4o-update/809376)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 2\ **Last updated:** [May 3, 2025, 11:30am UTC](https://community.deeplearning.ai/t/openai-rolls-back-sycophantic-gpt-4o-update/809376 "2025-05-03T11:30:45Z")
What are your thoughts on this? Reply below :down\_arrow: OpenAI reverted an April 25th update to GPT-4o that made the model excessively agreeable, particularly when validating users’ negative emotions. The update combin… --- ## [Nari Labs launches Dia, an open text-to-speech generator](https://community.deeplearning.ai/t/nari-labs-launches-dia-an-open-text-to-speech-generator/807745)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [April 29, 2025, 2:17pm UTC](https://community.deeplearning.ai/t/nari-labs-launches-dia-an-open-text-to-speech-generator/807745 "2025-04-29T14:17:12Z")
Subscribe for free access to :arrow\_forward: Data Points! Nari Labs, a two-person startup, released Dia, a 1.6 billion parameter text-to-speech model that generates naturalistic dialogue directly from text prompts. The … --- ## [OpenAI launches GPT-4.1 model family](https://community.deeplearning.ai/t/openai-launches-gpt-4-1-model-family/800663)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [April 14, 2025, 9:38pm UTC](https://community.deeplearning.ai/t/openai-launches-gpt-4-1-model-family/800663 "2025-04-14T21:38:02Z")
Subscribe for free access to :arrow\_forward: Data Points! OpenAI released three new models in its API: GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano, all outperforming previous versions with significant gains in coding and in… --- ## [OpenAI countersues Elon Musk, alleges harassment campaign](https://community.deeplearning.ai/t/openai-countersues-elon-musk-alleges-harassment-campaign/799277)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 1\ **Last updated:** [April 12, 2025, 6:39pm UTC](https://community.deeplearning.ai/t/openai-countersues-elon-musk-alleges-harassment-campaign/799277 "2025-04-12T18:39:28Z")
Subscribe for free access to :arrow\_forward: Data Points! OpenAI asked a federal judge to halt what it describes as a pattern of harassment and “unlawful and unfair action” by billionaire Elon Musk. OpenAI claims Musk, … --- ## [Meta releases Llama 4 models, claiming superior performance](https://community.deeplearning.ai/t/meta-releases-llama-4-models-claiming-superior-performance/797243)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [April 7, 2025, 8:21pm UTC](https://community.deeplearning.ai/t/meta-releases-llama-4-models-claiming-superior-performance/797243 "2025-04-07T20:21:05Z")
Subscribe for free access to :arrow\_forward: Data Points! Meta launched two new Llama 4 multimodal models, boasting performance improvements over previous generations and 10 million token context windows. Llama 4 Maveri… --- ## [Claude Sonnet 3.5-based agent tops new OpenAI ML research benchmark](https://community.deeplearning.ai/t/claude-sonnet-3-5-based-agent-tops-new-openai-ml-research-benchmark/795899)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [April 4, 2025, 9:08pm UTC](https://community.deeplearning.ai/t/claude-sonnet-3-5-based-agent-tops-new-openai-ml-research-benchmark/795899 "2025-04-04T21:08:01Z")
Subscribe for free access to :arrow\_forward: Data Points! OpenAI researchers released PaperBench, a new dataset and benchmark that tests large language models’ ability to recreate AI and machine learning papers from scr… --- ## [Language model mysteries revealed: How Claude thinks and plans](https://community.deeplearning.ai/t/language-model-mysteries-revealed-how-claude-thinks-and-plans/793862)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 1\ **Last updated:** [March 31, 2025, 7:09pm UTC](https://community.deeplearning.ai/t/language-model-mysteries-revealed-how-claude-thinks-and-plans/793862 "2025-03-31T19:09:24Z")
Subscribe for free access to :arrow\_forward: Data Points! Anthropic researchers used new interpretability techniques modeled on laboratory biology to examine how Claude processes information internally. By conducting ex… --- ## [Mistral releases Arabic-focused language model for Middle East market](https://community.deeplearning.ai/t/mistral-releases-arabic-focused-language-model-for-middle-east-market/771905)
**Author:** [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\ **Replies:** 0\ **Last updated:** [February 17, 2025, 9:10pm UTC](https://community.deeplearning.ai/t/mistral-releases-arabic-focused-language-model-for-middle-east-market/771905 "2025-02-17T21:10:05Z")
Subscribe for free access to :arrow\_forward: Data Points! French AI startup Mistral launched Mistral Saba, a 24-billion-parameter language model designed for Arabic-speaking countries. The model outperforms Mistral’s ge… [Next page](https://community.deeplearning.ai/tag/data-points/54.md?match_all_tags=true&page=1&tags%5B%5D=data-points)