{
  "name": "AI Model Releases Timeline",
  "description": "Frontier and open model releases as they happen, with what actually changed. Compiled from ANOTHER News daily coverage.",
  "url": "https://anothernews.io/trackers/model-releases/",
  "publisher": "ANOTHER News",
  "publisherUrl": "https://anothernews.io/",
  "license": "https://creativecommons.org/licenses/by/4.0/",
  "licenseName": "CC BY 4.0",
  "terms": "https://anothernews.io/data/",
  "citation": "ANOTHER News, AI Model Releases Timeline, https://anothernews.io/trackers/model-releases/",
  "updated": "2026-09-23",
  "schemaVersion": 2,
  "rowCount": 39,
  "rows": [
    {
      "date": "2026-06-16",
      "model": "GLM-5.2",
      "developer": "Z.ai",
      "kind": "release",
      "license": "open-weights",
      "params": "744B total, 40B active",
      "claim": "leading open-weights model on the Artificial Analysis Intelligence Index v4.1, scoring 51 against MiniMax-M3 at 44, DeepSeek V4 Pro at 44 and Kimi K2.6 at 43; same size as GLM-5.1 but 11 points higher (Artificial Analysis)",
      "event": "Z.ai releases GLM-5.2, which Artificial Analysis rates the leading open-weights model on its Intelligence Index v4.1 at 51.",
      "source": "https://artificialanalysis.ai/articles/glm-5-2-is-the-new-leading-open-weights-model-on-the-artificial-analysis-intelligence-index"
    },
    {
      "date": "2026-07-09",
      "model": "GPT-5.6 Sol, Terra and Luna",
      "developer": "OpenAI",
      "kind": "release",
      "license": "proprietary",
      "event": "GPT-5.6 Sol, Terra and Luna go public.",
      "source": "https://openai.com/index/gpt-5-6/"
    },
    {
      "date": "2026-07-16",
      "model": "Kimi K3",
      "developer": "Moonshot AI",
      "kind": "research-result",
      "event": "Kimi K3 takes the number one spot on the arena leaderboard from Claude and GPT.",
      "source": "https://x.com/arena/status/2077824029126504525"
    },
    {
      "date": "2026-07-17",
      "model": "Gemini 3.5 Pro",
      "developer": "Google",
      "kind": "delay",
      "event": "Google's Gemini 3.5 is reported delayed while headlines claim the opposite.",
      "source": "https://www.bloomberg.com/news/articles/2026-07-16/google-gemini-launch-delayed-as-tech-falls-short-of-internal-goals"
    },
    {
      "date": "2026-07-31",
      "model": "DeepSeek-V4-Flash-0731",
      "developer": "DeepSeek",
      "kind": "release",
      "claim": "Terminal Bench 2.1 at 82.7, up from 61.8, at unchanged prices (DeepSeek)",
      "event": "DeepSeek releases V4-Flash-0731, pushing Terminal Bench 2.1 from 61.8 to 82.7 at unchanged prices.",
      "source": "https://github.com/ggml-org/llama.cpp",
      "article": "https://anothernews.io/news/llamacpp-deepseek-v4-flash/"
    },
    {
      "date": "2026-08-03",
      "model": "Qwen3.8-Max",
      "developer": "Alibaba",
      "kind": "release",
      "params": "2.4T total, 95B active per token",
      "claim": "first Qwen-Max-class model to get open weights (Alibaba)",
      "event": "Alibaba announces Qwen3.8-Max, 2.4 trillion parameters, with open weights promised within a week.",
      "source": "https://github.com/QwenLM",
      "article": "https://anothernews.io/news/qwen38-max-open-weights/"
    },
    {
      "date": "2026-08-06",
      "model": "Muse Code",
      "developer": "Meta",
      "kind": "product",
      "availability": [
        "terminal",
        "macos",
        "linux"
      ],
      "event": "Meta launches Muse Code, with a cheap tier that trains on user prompts and code.",
      "source": "https://research.meta.ai/blog/introducing-muse-code-and-muse-spark-1-2",
      "article": "https://anothernews.io/news/meta-muse-code/"
    },
    {
      "date": "2026-08-10",
      "model": "Muse Glimmer",
      "developer": "Meta",
      "kind": "open-weights",
      "license": "Apache-2.0",
      "params": "30B",
      "availability": [
        "weights"
      ],
      "event": "Meta releases Muse Glimmer, a 30B open-weights model that runs on a single consumer GPU.",
      "source": "https://research.meta.ai/blog/introducing-muse-glimmer-open-agentic-model",
      "article": "https://anothernews.io/news/meta-muse-glimmer/"
    },
    {
      "date": "2026-08-10",
      "model": "WeatherNext",
      "developer": "Google DeepMind",
      "kind": "open-weights",
      "availability": [
        "weights",
        "code"
      ],
      "claim": "cyclone track and intensity a full day further out than prior systems (Nature paper, DeepMind)",
      "event": "DeepMind opens WeatherNext, which beats existing cyclone forecasts by a full day.",
      "source": "https://deepmind.google/blog/weathernext-ai-model-achieves-breakthrough-in-forecasting-cyclones/",
      "article": "https://anothernews.io/news/weathernext-open-source/"
    },
    {
      "date": "2026-08-11",
      "model": "GPT-5.6-Cyber",
      "developer": "OpenAI",
      "kind": "release",
      "license": "proprietary",
      "availability": [
        "restricted"
      ],
      "claim": "answers 95% of advanced security queries against about 2% for Daybreak Blue (OpenAI)",
      "event": "OpenAI ships GPT-5.6-Cyber, a security-specialized model that answers 95 percent of the questions its general models refuse.",
      "source": "https://x.com/OpenAI/status/2086864365379010729",
      "article": "https://anothernews.io/news/gpt56-cyber-daybreak/"
    },
    {
      "date": "2026-08-13",
      "model": "SL2T",
      "developer": "Google DeepMind",
      "kind": "release",
      "availability": [
        "device"
      ],
      "event": "DeepMind releases SL2T, a sign language translation model shipping in the Pixel 11 keyboard.",
      "source": "https://deepmind.google/blog/putting-sign-language-ai-into-users-hands/",
      "article": "https://anothernews.io/news/deepmind-sl2t-sign-language/"
    },
    {
      "date": "2026-08-14",
      "model": "DeepSeek-V4-Pro-0813",
      "developer": "DeepSeek",
      "kind": "price-change",
      "availability": [
        "api",
        "app"
      ],
      "claim": "peak output tokens from $0.87 to $3.96 per million, cache-hit input from $0.003625 to $0.044 (DeepSeek)",
      "event": "DeepSeek ships V4 Pro and raises API prices up to 12x with peak-hour billing.",
      "source": "https://api-docs.deepseek.com/quick_start/pricing",
      "article": "https://anothernews.io/news/deepseek-price-hike-12x/"
    },
    {
      "date": "2026-08-14",
      "model": "GPT-5.6 Sol on Ultrafast",
      "developer": "OpenAI",
      "kind": "preview",
      "availability": [
        "api"
      ],
      "claim": "up to 750 output tokens per second, up to 14x Standard, on Cerebras hardware (OpenAI)",
      "event": "OpenAI previews Ultrafast, running GPT-5.6 Sol at up to 750 tokens per second.",
      "source": "https://openai.com/index/previewing-ultrafast/",
      "article": "https://anothernews.io/news/openai-ultrafast-750tps/"
    },
    {
      "date": "2026-08-16",
      "model": "Gemini 3.7 Flash",
      "developer": "Google",
      "kind": "release",
      "license": "proprietary",
      "context_tokens": 1000000,
      "availability": [
        "api"
      ],
      "claim": "FrontierCode 43.6% (from 34.4%), DeepSWE 65.3% (from 49.0%); $0.75/$3.75 per million until 2027-01-01 (Google)",
      "event": "Google releases Gemini 3.7 Flash three weeks after 3.6, while Gemini 3.5 Pro stays delayed.",
      "source": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-3-7-flash/",
      "article": "https://anothernews.io/news/gemini-37-flash/"
    },
    {
      "date": "2026-08-16",
      "model": "MotionBricks",
      "developer": "Nvidia",
      "kind": "release",
      "claim": "one controller trained on 350,000 mocap clips, 15,000 FPS at 2 ms latency (Nvidia Research)",
      "event": "Nvidia's MotionBricks animates game characters and a humanoid robot from one model trained on 350,000 clips.",
      "source": "https://nvlabs.github.io/motionbricks/",
      "article": "https://anothernews.io/news/nvidia-motionbricks/"
    },
    {
      "date": "2026-08-18",
      "model": "Model 2",
      "developer": "Anthropic",
      "kind": "disclosure",
      "claim": "\"somewhat more capable than Mythos 5\", no plans to release externally (Anthropic risk report)",
      "event": "Anthropic's risk report discloses an unreleased Model 2 more capable than Mythos 5.",
      "source": "https://www.anthropic.com/aug-2026-risk-report",
      "article": "https://anothernews.io/news/anthropic-model-2-unreleased/"
    },
    {
      "date": "2026-08-19",
      "developer": "OpenAI",
      "kind": "pause",
      "claim": "RL training on newest deployment-bound models paused two weeks; largest frontier run on hold (OpenAI)",
      "event": "OpenAI pauses frontier RL training for two weeks and keeps its largest run on hold.",
      "source": "https://x.com/sama/status/2089787807611195475",
      "article": "https://anothernews.io/news/openai-stopped-training-run/"
    },
    {
      "date": "2026-08-20",
      "model": "Claude",
      "developer": "Anthropic",
      "kind": "research-result",
      "claim": "rank 30 then rank 31 elliptic curve, with Levent Alpoge and Ava Howell; Epoch AI marked the problem \"Solved (AI)\"",
      "event": "Claude, working with mathematicians, finds a rank 30 elliptic curve, the first improvement on the 2024 record; a second improvement follows within the week.",
      "source": "https://epoch.ai/frontiermath/open-problems/elliptic-curve-rank",
      "article": "https://anothernews.io/news/claude-math-record/"
    },
    {
      "date": "2026-08-22",
      "model": "Ox Alpha",
      "kind": "preview",
      "context_tokens": 1048576,
      "availability": [
        "api",
        "preview"
      ],
      "claim": "near 80% on a 10-task DeepSWE subset in one independent test; no lab attached",
      "event": "A model listed only as Ox Alpha appears on OpenRouter with no lab attached, a 1,048,576-token context window and near-unlimited free usage.",
      "source": "https://thenextweb.com/news/ox-alpha-stealth-model-openrouter-anonymous-provider",
      "article": "https://anothernews.io/news/ox-alpha-mystery-model/"
    },
    {
      "date": "2026-08-23",
      "model": "GLM-5.3",
      "developer": "Z.ai",
      "kind": "delay",
      "claim": "2,436 flaws across 269 open-source projects, 1,097 medium to high severity (Z.ai testing)",
      "event": "Z.ai delays GLM-5.3 open weights by two weeks after the model finds 2,436 flaws across 269 open-source projects, 1,097 of them medium to high severity.",
      "source": "https://venturebeat.com/technology/glm-5-3-is-here-with-advanced-cyber-capabilities-and-reportedly-already-found-a-serious-vulnerability-in-cursor",
      "article": "https://anothernews.io/news/zai-glm53-bugs-weights/"
    },
    {
      "date": "2026-08-24",
      "model": "GPT-5.6 Sol",
      "developer": "OpenAI",
      "kind": "price-change",
      "claim": "API and credit pricing cut over 20% for three months; $4/$20 per million, down from $5/$30, through 2026-11-21 (OpenAI)",
      "event": "OpenAI cuts GPT-5.6 Sol API and credit pricing by over 20 percent for three months.",
      "source": "https://openai.com/api/pricing/",
      "article": "https://anothernews.io/news/openai-sol-price-cut/"
    },
    {
      "date": "2026-08-26",
      "model": "Claude Cowork browser",
      "developer": "Anthropic",
      "kind": "product",
      "availability": [
        "macos",
        "windows"
      ],
      "event": "Anthropic ships a browser inside Claude Cowork: the desktop app opens a side panel and navigates, clicks and fills forms for Pro, Max and Team plans on macOS and Windows.",
      "source": "https://claude.com/blog/cowork-built-in-browser",
      "article": "https://anothernews.io/news/claude-cowork-browser/"
    },
    {
      "date": "2026-08-26",
      "model": "Qwen3.8-Flash-Next",
      "developer": "Alibaba",
      "kind": "open-weights",
      "params": "176B total (125B model plus 51B N-gram lookup), about 6B active per token",
      "context_tokens": 262144,
      "availability": [
        "weights"
      ],
      "event": "Alibaba open-weights Qwen3.8-Flash-Next: 176 billion parameters (125B model plus a 51B N-gram lookup) with only about 6B active per token.",
      "source": "https://qwen.ai/blog?id=qwen3.8-flash-next"
    },
    {
      "date": "2026-08-27",
      "model": "Instinct",
      "developer": "Instinct",
      "kind": "funding",
      "claim": "$250M Series B, $350M total at a $2.5B valuation, still private beta (TechCrunch)",
      "event": "Instinct, an AI assistant reached by text or phone call that acts through connected apps, is valued at $2.5 billion while still not publicly available.",
      "source": "https://techcrunch.com/2026/08/26/viral-ai-startup-instinct-has-raised-350-million-at-a-2-5-billion-valuation/",
      "article": "https://anothernews.io/news/instinct-assistant-25b/"
    },
    {
      "date": "2026-08-27",
      "model": "Model Hardware Standard",
      "developer": "Anthropic",
      "kind": "product",
      "availability": [
        "research-preview"
      ],
      "claim": "QuEra laser stabilization from 58% to 99.3% (Anthropic)",
      "event": "Anthropic opens the research preview of the Model Hardware Standard, letting agents discover and operate lab equipment; QuEra laser stabilization went from 58% to 99.3%.",
      "source": "https://www.anthropic.com/news/model-hardware-standard-research-preview",
      "article": "https://anothernews.io/news/anthropic-mhs/"
    },
    {
      "date": "2026-08-29",
      "model": "MiniMax H3 Max",
      "developer": "fal",
      "kind": "release",
      "availability": [
        "api"
      ],
      "claim": "5-second 768p clip with audio in under 3 seconds, about 35x the official MiniMax endpoint (fal)",
      "event": "fal ships MiniMax H3 Max, rendering 5-second audio-video clips in under 3 seconds, faster than playback.",
      "source": "https://fal.ai/minimax-h3-max",
      "article": "https://anothernews.io/news/infinite-slop-livestream/"
    },
    {
      "date": "2026-08-30",
      "model": "DALL-E GPT",
      "developer": "OpenAI",
      "kind": "withdrawal",
      "event": "OpenAI retires the original DALL-E GPT from ChatGPT; ChatGPT Images takes over.",
      "source": "https://help.openai.com/en/articles/6825453-chatgpt-release-notes",
      "article": "https://anothernews.io/news/dalle-retires-today/"
    },
    {
      "date": "2026-08-30",
      "model": "GPT-5.6",
      "developer": "OpenAI",
      "kind": "research-result",
      "claim": "improved the 2018 Ford-Green-Konyagin-Maynard-Tao bound on large prime gaps, reported by Jared Lichtman; no proof check yet",
      "event": "GPT-5.6 breaks the 2018 record on large gaps between primes set by Ford, Green, Konyagin, Maynard and Tao, per Oxford number theorist Jared Lichtman.",
      "source": "https://x.com/jdlichtman/status/2094040463443673227",
      "article": "https://anothernews.io/news/gpt56-prime-gap-record/"
    },
    {
      "date": "2026-09-01",
      "model": "CameraJet",
      "developer": "Dyson",
      "kind": "product",
      "claim": "$499 device, 100,000-pixel camera at 28 images per second, ML trained on about 470,000 dental images (Dyson)",
      "event": "Dyson launches CameraJet, its first toothbrush: a $499 device with a 100,000-pixel camera at 28fps and AI that fires mouth rinse at gaps between teeth.",
      "source": "https://www.bloomberg.com/news/articles/2026-09-01/dyson-announces-499-camerajet-smart-toothbrush-with-ai-built-in-camera",
      "article": "https://anothernews.io/news/dyson-camerajet-toothbrush/"
    },
    {
      "date": "2026-09-01",
      "model": "Claude Fable 5.1 and Mythos 5.1",
      "developer": "Anthropic",
      "kind": "release",
      "license": "proprietary",
      "availability": [
        "api",
        "app",
        "aws",
        "gcp",
        "azure"
      ],
      "claim": "Fable 5.1 at 55.8% on Terminal-Bench 4.0 against 42.0% for Fable 5; cache reads 75% cheaper (Anthropic)",
      "event": "Anthropic launches Claude Fable 5.1 and Mythos 5.1: 55.8% on Terminal-Bench 4.0, cache reads 75% cheaper; Fable models run on usage credits for Pro subscribers.",
      "source": "https://www.anthropic.com/claude-fable-and-mythos-5-1",
      "article": "https://anothernews.io/news/fable-51-not-in-pro/"
    },
    {
      "date": "2026-09-01",
      "model": "GPT-6 Astra",
      "developer": "OpenAI",
      "kind": "designation",
      "claim": "first model designated at the Critical cybersecurity threshold of OpenAI's Preparedness Framework (OpenAI)",
      "event": "OpenAI designates Astra the first model at its Critical cybersecurity threshold and says it will release it with limited cyber access via Daybreak Blue.",
      "source": "https://openai.com/index/path-to-astra/",
      "article": "https://anothernews.io/news/openai-astra-critical/"
    },
    {
      "date": "2026-09-02",
      "model": "Gemini 3.8 Flash",
      "developer": "Google",
      "kind": "release",
      "event": "Google releases Gemini 3.8 Flash, third Flash in six weeks, ahead of Opus 5 on Google's tables.",
      "source": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/3-8-flash-and-3-8-flash-cyber/",
      "article": "https://anothernews.io/news/gemini-38-flash-six-weeks/"
    },
    {
      "date": "2026-09-02",
      "model": "Muse Spark 1.3",
      "developer": "Meta",
      "kind": "release",
      "claim": "75.4% on DeepSWE v1.1 (Meta)",
      "event": "Meta releases Muse Spark 1.3, reporting 75.4% on DeepSWE v1.1.",
      "source": "https://www.neowin.net/news/meta-rolls-out-muse-spark-13-with-stronger-coding-and-agentic-performance/",
      "article": "https://anothernews.io/news/muse-spark-13-deepswe/"
    },
    {
      "date": "2026-09-03",
      "model": "GPT-6 Astra",
      "developer": "OpenAI",
      "kind": "release",
      "license": "proprietary",
      "availability": [
        "api",
        "app",
        "azure",
        "aws"
      ],
      "claim": "99.9% on ARC-AGI-3 and 100% on ExploitBench (OpenAI)",
      "event": "OpenAI launches GPT-6 Astra, scoring 99.9% on ARC-AGI-3 and 100% on ExploitBench; rollout to all paid ChatGPT tiers, API, Azure and AWS Bedrock within days.",
      "source": "https://openai.com/index/gpt-6-astra/",
      "article": "https://anothernews.io/news/gpt6-astra-launch/"
    },
    {
      "date": "2026-09-05",
      "model": "WeatherNext 3",
      "developer": "Google DeepMind",
      "kind": "release",
      "availability": [
        "search",
        "maps",
        "gemini"
      ],
      "claim": "first global weather model with hourly forecasts, 5km resolution against a 25km grid (Google DeepMind)",
      "event": "Google DeepMind launches WeatherNext 3, the first global weather model to produce a new forecast every hour from live satellite data, at 5 km resolution; rolling into Search, Maps and Gemini.",
      "source": "https://blog.google/innovation-and-ai/models-and-research/google-deepmind/introducing-weathernext-3/",
      "article": "https://anothernews.io/news/wym-weathernext-hourly/"
    },
    {
      "date": "2026-09-08",
      "model": "Muse",
      "developer": "Meta",
      "kind": "product",
      "availability": [
        "ios",
        "android",
        "web"
      ],
      "event": "Meta launches Muse, a personal AI agent that reads email, books travel, negotiates and pays with the user's saved card; US on iOS, Android and web, free tier plus $20 and $100 plans.",
      "source": "https://www.axios.com/2026/09/08/meta-debuts-muse-personal-ai-agent",
      "article": "https://anothernews.io/news/meta-muse-agent/"
    },
    {
      "date": "2026-09-12",
      "model": "GPT-6 Astra",
      "developer": "OpenAI",
      "kind": "incident",
      "event": "OpenAI fixes three problems behind GPT-6 Astra's quality drop (old skills blocking self-checks, an opt-in context experiment hitting about 4,000 to 5,000 users, misconfigured engines) and resets usage limits.",
      "source": "https://x.com/thsottiaux/status/2098612714704891959",
      "article": "https://anothernews.io/news/astra-three-bugs/"
    },
    {
      "date": "2026-09-22",
      "model": "Claude Opus 5.5",
      "developer": "Anthropic",
      "kind": "release",
      "availability": [
        "app",
        "api",
        "aws",
        "gcp",
        "azure"
      ],
      "claim": "40% less than Opus 5 on typical workloads (Anthropic)",
      "event": "Anthropic releases Claude Opus 5.5 at $4/$20 per million tokens, 20% below Opus 5, raises five-hour usage limits on paid plans and gives subscribers a limit reset.",
      "source": "https://www.anthropic.com/claude-opus-5-5",
      "article": "https://anothernews.io/news/claude-opus-55/"
    },
    {
      "date": "2026-09-22",
      "model": "GPT-6 Sol and GPT-6 Luna",
      "developer": "OpenAI",
      "kind": "release",
      "availability": [
        "api"
      ],
      "event": "OpenAI releases GPT-6 Sol ($2/$10 per million tokens) and GPT-6 Luna ($0.10/$0.50), half the GPT-5.6 prices, and tells VentureBeat the prices are permanent.",
      "source": "https://openai.com/index/introducing-gpt-6-sol-and-luna/",
      "article": "https://anothernews.io/news/openai-half-price/"
    }
  ]
}