Upload updates.json with huggingface_hub

updates.json

@@ -0,0 +1,126 @@
+{
+  "lastUpdated": "2026-07-08T10:57:03.336411Z",
+  "updates": [
+    {
+      "type": "launch",
+      "tag_emoji": "🚀",
+      "tag_label": "New Launch",
+      "tag_color": "#3b82f6",
+      "title": "Claude Science — Anthropic's research workbench",
+      "desc": "Anthropic launched Claude Science, a workbench with 60+ preconfigured tools for pharmaceutical and academic researchers, plus a new internal drug-discovery program targeting neglected diseases.",
+      "date": "Jul 5, 2026",
+      "tool_name": "Claude Science",
+      "url": "https://www.anthropic.com/"
+    },
+    {
+      "type": "update",
+      "tag_emoji": "🔄",
+      "tag_label": "Update",
+      "tag_color": "#f59e0b",
+      "title": "Claude Fable 5 restored after 19-day government ban",
+      "desc": "Anthropic redeployed Claude Fable 5 globally after the US Commerce Department lifted the export-control order that had pulled the frontier model offline on June 12.",
+      "date": "Jul 1, 2026",
+      "tool_name": "Claude Fable 5",
+      "url": "https://www.anthropic.com/"
+    },
+    {
+      "type": "model",
+      "tag_emoji": "🧠",
+      "tag_label": "New Model",
+      "tag_color": "#a855f7",
+      "title": "Amália — Portugal's first open Portuguese LLM",
+      "desc": "The Portuguese government unveiled Amália, an open-source 9B-parameter language model in European Portuguese, published under Apache 2.0 by a consortium of Portuguese universities.",
+      "date": "Jul 1, 2026",
+      "tool_name": "Amália",
+      "url": "https://huggingface.co/amalia-llm"
+    },
+    {
+      "type": "model",
+      "tag_emoji": "🧠",
+      "tag_label": "New Model",
+      "tag_color": "#a855f7",
+      "title": "Claude Sonnet 5 — Anthropic's most agentic Sonnet",
+      "desc": "Anthropic launched Claude Sonnet 5 as the new default for Free and Pro users, scoring 63.2% on SWE-Bench Pro with a native 1M-token context at $2/$10 introductory pricing.",
+      "date": "Jun 30, 2026",
+      "tool_name": "Claude Sonnet 5",
+      "url": "https://www.anthropic.com/news/claude-sonnet-5"
+    },
+    {
+      "type": "model",
+      "tag_emoji": "🧠",
+      "tag_label": "New Model",
+      "tag_color": "#a855f7",
+      "title": "LongCat-2.0 — Meituan's 1.6T open-source coding model",
+      "desc": "Meituan open-sourced LongCat-2.0, a 1.6T-parameter MoE agentic coding model trained entirely on Chinese chips that hit 59.5% on SWE-bench Pro under an MIT license.",
+      "date": "Jun 30, 2026",
+      "tool_name": "LongCat-2.0",
+      "url": "https://www.longcatai.org/"
+    },
+    {
+      "type": "model",
+      "tag_emoji": "🧠",
+      "tag_label": "New Model",
+      "tag_color": "#a855f7",
+      "title": "Gemini 3.1 Flash Image & Gemini 3 Pro Image",
+      "desc": "Google released two new image-generation models through the Gemini API and AI Studio, offering unified multimodal endpoints for high-volume ($0.50/$3) and quality-first ($2/$12) creative workflows.",
+      "date": "Jun 30, 2026",
+      "tool_name": "Gemini 3.1 Flash Image",
+      "url": "https://ai.google.dev/"
+    },
+    {
+      "type": "model",
+      "tag_emoji": "🧠",
+      "tag_label": "New Model",
+      "tag_color": "#a855f7",
+      "title": "GPT-5.6 Sol, Terra & Luna — OpenAI's new family",
+      "desc": "OpenAI previewed the GPT-5.6 family with new state-of-the-art on Terminal-Bench 2.1, launched under a government-gated access list to roughly 20 trusted partners before broad GA.",
+      "date": "Jun 26, 2026",
+      "tool_name": "GPT-5.6",
+      "url": "https://openai.com/index/previewing-gpt-5-6-sol/"
+    },
+    {
+      "type": "update",
+      "tag_emoji": "🔄",
+      "tag_label": "Update",
+      "tag_color": "#f59e0b",
+      "title": "gpt-realtime-2.1 — 25% lower voice latency",
+      "desc": "OpenAI released gpt-realtime-2.1 and gpt-realtime-2.1-mini for low-latency voice and multimodal agents, cutting p95 latency by at least 25% with improved noise handling and tool use.",
+      "date": "Jun 24, 2026",
+      "tool_name": "gpt-realtime-2.1",
+      "url": "https://platform.openai.com/docs/models"
+    },
+    {
+      "type": "research",
+      "tag_emoji": "🔬",
+      "tag_label": "Research",
+      "tag_color": "#10b981",
+      "title": "FrontierMath v2 — Epoch's harder math benchmark",
+      "desc": "Epoch AI released FrontierMath v2 with corrected and removed problematic items, featuring a Tier 4 set of the hardest private research-level problems for evaluating frontier reasoning.",
+      "date": "Jun 12, 2026",
+      "tool_name": "FrontierMath v2",
+      "url": "https://epoch.ai/frontiermath"
+    },
+    {
+      "type": "research",
+      "tag_emoji": "🔬",
+      "tag_label": "Research",
+      "tag_color": "#10b981",
+      "title": "GeneBench-Pro — OpenAI's computational biology benchmark",
+      "desc": "OpenAI introduced GeneBench-Pro, a research-level benchmark for judging AI agents on long-horizon genomics and quantitative-biology analyses.",
+      "date": "Jun 22, 2026",
+      "tool_name": "GeneBench-Pro",
+      "url": "https://openai.com/research"
+    },
+    {
+      "type": "update",
+      "tag_emoji": "🔄",
+      "tag_label": "Update",
+      "tag_color": "#f59e0b",
+      "title": "GPT-5.5 Instant Mini rolls out in ChatGPT",
+      "desc": "OpenAI rolled out GPT-5.5 Instant Mini as the new ChatGPT fallback model, delivering better intent tracking, tone calibration, and fewer factual issues than GPT-5.3 Instant Mini.",
+      "date": "Jun 20, 2026",
+      "tool_name": "ChatGPT",
+      "url": "https://help.openai.com/"
+    }
+  ]
+}