Release Tracker
Key milestones in AI history · newest first
starGemini 2.5 Pro Deep Think
llmGoogle's most capable reasoning model ever. Dropped June 22 — leads multiple benchmarks.
starClaude Fable 5
llmAnthropic's new frontier tier above Opus. 1M context, 80.3% SWE-bench Pro, $10/$50 per 1M. Free on Pro/Max through June 22.
Gemini 3.5 Pro (preview)
multimodalAnnounced at Google I/O 2026. Limited Vertex preview, GA imminent. Powers Google AI Mode in Search.
Siri AI (Apple WWDC 2026)
llmApple rebuilt Siri with generative AI powered by Google Gemini models. Announced at WWDC June 8.
Microsoft MAI-Thinking-1
llmMicrosoft's first in-house reasoning model family, reducing strategic dependence on OpenAI.
starGemini 3.5 Flash
llmFast, multimodal, 1M context. Powers Google AI Mode in Search. Announced at I/O 2026 May 19.
starGPT-5.5
llmReleased April 23. $5/$30 per 1M tokens, ~1M context. Neck-and-neck with Claude Opus on coding.
Grok 4.3
llmxAI flagship released April 30. $1.25/$2.50 per 1M — frontier quality at mid-tier price. 1M context, native video input.
starDeepSeek V4 Pro
llm1.6T parameter MoE, MIT license, 1M context, $1.74/$3.48 per 1M. First OSS model at near-frontier level.
starGemini 3.1 Pro
multimodalReleased Feb 19. Tops 13/16 benchmarks: 94.3% GPQA Diamond, 80.6% SWE-bench. $2/$12 — 7.5x cheaper than GPT-5.4.
Qwen 3.5 Family
llmOpen-weight 235B and smaller models. Strongest OSS reasoning — leads GPQA Diamond and AIME. 262K native context.
DeepSeek V3.2
llmFinal V3 iteration — MIT-licensed MoE model matching GPT-4 class at <$0.30/1M input. Massive OSS milestone.
Grok-2.5 (open weights)
llmxAI open-sourced Grok-2.5 weights — strong performance for an open model at this scale.
starLlama 4 (Scout + Maverick)
llmScout: 10M context at 2600 t/s. Maverick: 109B MoE, 17B active, 85.5% MMLU. New OSS baseline.