[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news:stories\u002F32:en:{}":3},{"story":4,"neighbours":45,"timeline":66,"suggestions":119,"indexable":120,"locale":121},{"id":5,"path":6,"slug":7,"headline":8,"summary":9,"summary_html":10,"why_it_matters":11,"section":12,"section_label":13,"quote":14,"author_name":14,"author_handle":14,"kind":15,"kind_label":16,"legal_status":14,"region":14,"importance":17,"source_name":18,"source_url":19,"source_published_at":20,"parent_id":14,"topics":21,"edition":30,"event":14,"created_at":33,"updated_at":33,"coverage":34},32,"\u002Fstories\u002F32-thinking-machines-releases-inkling-2-mini-an-open-model-that-runs-on-a","32-thinking-machines-releases-inkling-2-mini-an-open-model-that-runs-on-a","Thinking Machines releases Inkling 2 Mini, an open model that runs on a laptop","Thinking Machines Lab published Inkling 2 Mini, a 26.4B-parameter mixture-of-experts model with open weights under Apache 2.0, aimed at coding and agent work with a memory footprint of about 20GB. It is already available through Ollama, vLLM, SGLang and quantised GGUF builds.","\u003Cp>Thinking Machines Lab published Inkling 2 Mini, a 26.4B-parameter mixture-of-experts model with open weights under Apache 2.0, aimed at coding and agent work with a memory footprint of about 20GB. It is already available through Ollama, vLLM, SGLang and quantised GGUF builds.\u003C\u002Fp>","Teams that cannot send code or customer data to a US cloud API now have another capable coding model they can run on their own hardware; worth a test against your current assistant.","ai","AI & automation",null,"release","Product release",3,"Thinking Machines Lab","https:\u002F\u002Fthinkingmachines.ai\u002Fnews\u002Finkling-2-mini\u002F?utm_source=news.codingdelta.com&utm_medium=web","2026-11-20T00:00:00+01:00",[22,26],{"slug":23,"name":24,"path":25,"items_count":14},"open-source","Open source","\u002Ftopics\u002Fopen-source",{"slug":27,"name":28,"path":29,"items_count":14},"ai-models","AI models","\u002Ftopics\u002Fai-models",{"slug":31,"date":31,"path":32},"2026-11-21","\u002Feditions\u002F2026-11-21","2026-11-21T12:47:23+01:00",[35,38,41],{"source_name":36,"url":37,"title":14,"published_at":20},"X (@thinkymachines)","https:\u002F\u002Fx.com\u002Fthinkymachines\u002Fstatus\u002F2123871643797995872?utm_source=news.codingdelta.com&utm_medium=web",{"source_name":39,"url":40,"title":14,"published_at":20},"X (@ollama)","https:\u002F\u002Fx.com\u002Follama\u002Fstatus\u002F2123889582303281557?utm_source=news.codingdelta.com&utm_medium=web",{"source_name":42,"url":43,"title":44,"published_at":14},"Hugging Face","https:\u002F\u002Fhuggingface.co\u002Fthinkingmachines\u002FInkling-2-Mini?utm_source=news.codingdelta.com&utm_medium=web","thinkingmachines\u002FInkling-2-Mini",[46],{"relation":47,"label":48,"note":49,"item":50},"related","Related","Another smaller, cheaper model aimed at coding and agents",{"id":51,"path":52,"slug":53,"headline":54,"summary":55,"summary_html":56,"why_it_matters":57,"section":12,"section_label":13,"quote":14,"author_name":14,"author_handle":14,"kind":15,"kind_label":16,"legal_status":14,"region":14,"importance":17,"source_name":58,"source_url":59,"source_published_at":60,"parent_id":14,"topics":61,"edition":62,"event":14,"created_at":65,"updated_at":65},18,"\u002Fstories\u002F18-google-launches-gemini-4-neon-a-cheaper-mid-tier-model-for-coding-and","18-google-launches-gemini-4-neon-a-cheaper-mid-tier-model-for-coding-and","Google launches Gemini 4 Neon, a cheaper mid-tier model for coding and agents","Google released Gemini 4 Neon, sitting between Gemini 4 Hydrogen and Carbon, with a 1 million-token context window, four thinking levels and pricing of $3 per million input and $15 per million output tokens. It is available via the Gemini API, Google AI Studio (paid tier), Antigravity and Gemini for business.","\u003Cp>Google released Gemini 4 Neon, sitting between Gemini 4 Hydrogen and Carbon, with a 1 million-token context window, four thinking levels and pricing of $3 per million input and $15 per million output tokens. It is available via the Gemini API, Google AI Studio (paid tier), Antigravity and Gemini for business.\u003C\u002Fp>","Teams building on Gemini can test Neon as a lower-cost alternative to Carbon for coding and document-heavy automation.","Google","https:\u002F\u002Fblog.google\u002Finnovation-and-ai\u002Fmodels-and-research\u002Fgemini-models\u002Fintroducing-gemini-4-neon\u002F?utm_source=news.codingdelta.com&utm_medium=web","2026-11-18T00:00:00+01:00",[],{"slug":63,"date":63,"path":64},"2026-11-20","\u002Feditions\u002F2026-11-20","2026-11-20T11:38:49+01:00",{"items":67,"edges":114},[68,85,105,109],{"id":69,"path":70,"slug":71,"headline":72,"summary":73,"summary_html":74,"why_it_matters":75,"section":12,"section_label":13,"quote":14,"author_name":14,"author_handle":14,"kind":15,"kind_label":16,"legal_status":14,"region":14,"importance":17,"source_name":76,"source_url":77,"source_published_at":78,"parent_id":14,"topics":79,"edition":81,"event":14,"created_at":84,"updated_at":84},20,"\u002Fstories\u002F20-google-announces-gemini-4-carbon-its-new-frontier-model","20-google-announces-gemini-4-carbon-its-new-frontier-model","Google announces Gemini 4 Carbon, its new frontier model","Google announced Gemini 4 Carbon on 28 October as its most capable model for complex software engineering, reasoning and cybersecurity work. Early access goes to cyber defenders in Google's Fairwind Program, with broader developer and business access to follow; introductory API pricing was reported at $2 per million input and $10 per million output tokens.","\u003Cp>Google announced Gemini 4 Carbon on 28 October as its most capable model for complex software engineering, reasoning and cybersecurity work. Early access goes to cyber defenders in Google's Fairwind Program, with broader developer and business access to follow; introductory API pricing was reported at $2 per million input and $10 per million output tokens.\u003C\u002Fp>","Carbon sets the top of the Gemini 4 line-up; teams on Gemini can plan tests once it opens up, but there is no broad access yet.","9to5Google","https:\u002F\u002F9to5google.com\u002F2026\u002F10\u002F28\u002Fgemini-4-carbon\u002F?utm_source=news.codingdelta.com&utm_medium=web","2026-10-28T00:00:00+01:00",[80],{"slug":27,"name":28,"path":29,"items_count":14},{"slug":82,"date":82,"path":83},"2026-10-29","\u002Feditions\u002F2026-10-29","2026-11-20T12:51:27+01:00",{"id":86,"path":87,"slug":88,"headline":89,"summary":90,"summary_html":91,"why_it_matters":92,"section":12,"section_label":13,"quote":14,"author_name":14,"author_handle":14,"kind":15,"kind_label":16,"legal_status":14,"region":14,"importance":17,"source_name":58,"source_url":93,"source_published_at":94,"parent_id":14,"topics":95,"edition":101,"event":14,"created_at":104,"updated_at":104},24,"\u002Fstories\u002F24-google-releases-gemini-4-hydrogen-for-fast-low-cost-coding-and-agent-w","24-google-releases-gemini-4-hydrogen-for-fast-low-cost-coding-and-agent-w","Google releases Gemini 4 Hydrogen for fast, low-cost coding and agent work","Google released Gemini 4 Hydrogen on 12 November, a fast model for coding and agentic workflows that sits below Carbon in the Gemini 4 line-up. It is available in Google AI Studio and the Gemini API at $1.50 per million input and $7.50 per million output tokens; Google suggests pairing it with Carbon, with Carbon coordinating and Hydrogen doing the work.","\u003Cp>Google released Gemini 4 Hydrogen on 12 November, a fast model for coding and agentic workflows that sits below Carbon in the Gemini 4 line-up. It is available in Google AI Studio and the Gemini API at $1.50 per million input and $7.50 per million output tokens; Google suggests pairing it with Carbon, with Carbon coordinating and Hydrogen doing the work.\u003C\u002Fp>","For high-volume automations a cheaper, faster model can cut costs sharply; test Hydrogen on your own tasks before relying on Google's benchmark claims.","https:\u002F\u002Fblog.google\u002Finnovation-and-ai\u002Fmodels-and-research\u002Fgemini-models\u002Fintroducing-gemini-4-hydrogen\u002F?utm_source=news.codingdelta.com&utm_medium=web","2026-11-12T00:00:00+01:00",[96,97],{"slug":27,"name":28,"path":29,"items_count":14},{"slug":98,"name":99,"path":100,"items_count":14},"ai-agents","AI agents","\u002Ftopics\u002Fai-agents",{"slug":102,"date":102,"path":103},"2026-11-13","\u002Feditions\u002F2026-11-13","2026-11-20T12:51:29+01:00",{"id":51,"path":52,"slug":53,"headline":54,"summary":55,"summary_html":56,"why_it_matters":57,"section":12,"section_label":13,"quote":14,"author_name":14,"author_handle":14,"kind":15,"kind_label":16,"legal_status":14,"region":14,"importance":17,"source_name":58,"source_url":59,"source_published_at":60,"parent_id":14,"topics":106,"edition":108,"event":14,"created_at":65,"updated_at":65},[107],{"slug":27,"name":28,"path":29,"items_count":14},{"slug":63,"date":63,"path":64},{"id":5,"path":6,"slug":7,"headline":8,"summary":9,"summary_html":10,"why_it_matters":11,"section":12,"section_label":13,"quote":14,"author_name":14,"author_handle":14,"kind":15,"kind_label":16,"legal_status":14,"region":14,"importance":17,"source_name":18,"source_url":19,"source_published_at":20,"parent_id":14,"topics":110,"edition":113,"event":14,"created_at":33,"updated_at":33},[111,112],{"slug":23,"name":24,"path":25,"items_count":14},{"slug":27,"name":28,"path":29,"items_count":14},{"slug":31,"date":31,"path":32},[115,116,118],{"from":5,"to":51,"relation":47},{"from":51,"to":86,"relation":117},"follow_up",{"from":86,"to":69,"relation":117},[],true,"en"]