wiki

Every thing, typed and dated

One record per model, lab, tool, concept, technique, benchmark, dataset, paper or event. Every fact carries its source and the date it was read.

Sorted by name, A to Z. Nothing on this site is ordered by payment.

AI winterconceptactive
AionLabsorgactive
AionLabs: Aion 3.5modelactive
AionLabs: Aion 3.5 Minimodelactive
Alibaba Cloudorgactive
AllenAI: Olmo 3 32B Thinkmodelretired
AlphaFold at CASP14eventdead
AlphaGo versus Lee Sedoleventdead
Amazonorgactive
Anthropicorgactive
Anthropic: Claude Fable 5modelactive
Anthropic: Claude Fable 5.1modelactive
Anthropic: Claude Haiku 4.5modelactive
Anthropic: Claude Opus 4modelactive
Anthropic: Claude Opus 4.7modelactive
Anthropic: Claude Opus 4.7 (Fast)modelactive
Anthropic: Claude Opus 4.8modelactive
Anthropic: Claude Opus 4.8 (Fast)modelactive
Anthropic: Claude Opus 5.5modelactive
Anthropic: Claude Opus 5.5 (batch)modelactive
Arcee AI: Virtuoso Largemodelretired
Attention Is All You Needeventdead
Baidu: ERNIE 4.5 VL 424B A47B modeldeprecated
ByteDance Seedorgactive
ByteDance Seed: Seed 1.6modeldeprecated
ByteDance Seed: Seed 1.6 Flashmodeldeprecated
ByteDance Seed: Seed-2.0-Codemodeldeprecated
Chain-of-thought faithfulnessconceptactive
Chain-of-thought promptingtechniqueactive
Claude Opus 5modelactive
Claude Opus 5 (Fast)modelactive
Cohereorgactive
Cohere: Command Amodelactive
Cohere: Command A+modelactive
Covered Modelsconceptactive
Dartmouth Summer Research Project on Artificial Intelligenceeventdead
Deep Blue versus Kasparoveventdead
DeepSeekorgactive
DeepSeek: DeepSeek Flash Latestmodelactive
DeepSeek: DeepSeek Pro Latestmodelactive
DeepSeek: DeepSeek V3.1 Terminusmodeldeprecated
DeepSeek: DeepSeek V3.2modeldeprecated
DeepSeek: DeepSeek V3.2 Expmodeldeprecated
DeepSeek: DeepSeek V4 Flash 0731modelactive
DeepSeek: DeepSeek V4 Flash 0731 (batch)modelactive
DeepSeek: DeepSeek V4 Flash Vision Exp (batch)modelretired
DeepSeek: DeepSeek V4 Pro 0813 (batch)modelactive
DeepSeek: DeepSeek V4.1 Flashmodelactive
DeepSeek: DeepSeek V4.1 Flash (batch)modelactive
DeepSeek: R1 Distill Llama 70Bmodeldeprecated
Direct Preference Optimizationtechniqueactive
Distillationconceptactive
Dots Studio: Dots3-Note Preview (free)modeldeprecated
Effective context lengthconceptactive
ELIZAeventdead
Embeddingsconceptactive
Emergent abilitiesconceptactive
Fireworks: Ember-1modelactive
FlashAttentiontechniqueactive
Google DeepMindorgactive
Google: Gemini 2.5 Flashmodeldeprecated
Google: Gemini 2.5 Flash (batch)modeldeprecated
Google: Gemini 2.5 Flash Litemodeldeprecated
Google: Gemini 2.5 Promodeldeprecated
Google: Gemini 2.5 Pro (batch)modeldeprecated
Google: Gemini 2.5 Pro Preview 05-06modelretired
Google: Gemini 3.1 Pro Previewmodelactive
Google: Gemini 3.7 Flashmodelactive
Google: Gemini 3.8 Flashmodelactive
Google: Gemma 4 31B (batch)modelactive
Google: Nano Banana (Gemini 2.5 Flash Image)modeldeprecated
Grokkingconceptactive
Hallucinationconceptactive
IBM: Granite 4.1 8Bmodelactive
ILSVRC 2012eventdead
In-context learningconceptactive
Inception Labsorgactive
Inception: Mercury 2.5modelactive
Inception: Mercury 2.5 Previewmodelactive
inclusionAIorgactive
inclusionAI: Ling 3.0 Flash VLmodelactive
inclusionAI: Ling 3.0 Flash VL (free)modelactive
Inference.net: Schematron V2 Smallmodelactive
Inference.net: Schematron V2 Turbomodelactive
KV cacheconceptactive
Kwaipilot: KAT-Coder-Air V2.5modelretired
Kwaipilot: KAT-Coder-Pro V2modelactive
llama.cpptoolactive
Low-Rank Adaptationtechniqueactive
Meta Superintelligence Labsorgactive
Meta: Muse Glimmer 30Bmodelactive
Meta: Muse Glimmer 30B (batch)modelactive
Meta: Muse Spark 1.1modelactive
Meta: Muse Spark 1.2modelactive
Meta: Muse Spark 1.3modelactive
Meta: Muse Spark 1.3 Contributormodelactive
Microsoftorgactive
MiniMaxorgactive
MiniMax: MiniMax M2.1modeldeprecated
MiniMax: MiniMax M2.7 (free)modelactive
MiniMax: MiniMax M3 (batch)modelactive
MiniMax: MiniMax M3 (free)modelactive
Mistral AIorgactive
Mistral: Codestral 2508 (batch)modelactive
Mistral: Devstral 2 2512modelretired
Mistral: Ministral 3 8B 2512 (batch)modelactive
Mistral: Mistral Large 3 2512modelactive
Mistral: Mistral Large 3 2512 (batch)modelactive
Mistral: Mistral Medium 3.1 (batch)modelactive
Mistral: Mistral Small 4 (batch)modelactive
Mixture of Expertstechniqueactive
Model collapseconceptactive
Model Context Protocolconceptactive
Moonshot AIorgactive
MoonshotAI: Kimi K2.5modeldeprecated
MoonshotAI: Kimi K3modelactive
Nex AGI: Nex-N2-Minimodelactive
Nex AGI: Nex-N2-Promodelactive
Nex AGI: Nex-N2.5-Minimodelactive
Nex AGI: Nex-N2.5-Mini (free)modeldeprecated
Nex AGI: Nex-N2.5-Promodelactive
Nex AGI: Nex-N2.5-Pro (free)modeldeprecated
Nous Researchorgactive
Nous: Hermes 4 70Bmodelactive
NVIDIAorgactive
NVIDIA: Nemotron 3 Ultra (batch)modelactive
OpenAIorgactive
OpenAI Daybreakconceptactive
OpenAI GPT Astra Latestmodelactive
OpenAI GPT Latestmodelactive
OpenAI GPT Luna Latestmodelactive
OpenAI GPT Sol Latestmodelactive
OpenAI GPT Terra Latestmodelactive
OpenAI: GPT-4 Turbo Previewmodelretired
OpenAI: GPT-5.4modelactive
OpenAI: GPT-5.5modelactive
OpenAI: GPT-5.6 Lunamodelactive
OpenAI: GPT-5.6 Solmodelactive
OpenAI: GPT-5.6 Terramodelactive
OpenAI: GPT-6 Lunamodelactive
OpenAI: GPT-6 Luna (batch)modelactive
OpenAI: GPT-6 Luna Promodelactive
OpenAI: GPT-6 Luna Pro (batch)modelactive
OpenAI: GPT-6 Solmodelactive
OpenAI: GPT-6 Sol (batch)modelactive
OpenAI: GPT-6 Sol Promodelactive
OpenAI: GPT-6 Sol Pro (batch)modelactive
OpenAI: gpt-oss-20b (batch)modelactive
Paretomodelactive
Perplexity AIorgactive
Poolsideorgactive
PrismML: Ternary Bonsai 2 27Bmodelactive
Prompt tuningtechniqueactive
Proximal Policy Optimizationtechniqueactive
Pruningtechniqueactive
Quantizationtechniqueactive
Qwen: Qwen3.5-9B (batch)modelactive
Qwen: Qwen3.8 2.4T A95B (batch)modelactive
Qwen: Qwen3.8 27Bmodelactive
Qwen: Qwen3.8 27B (free)modelactive
Qwen: Qwen3.8 Maxmodelactive
Qwen: Qwen3.8 Max Primemodelactive
Qwen: Qwen3.8 Omni Flashmodelactive
Reinforcement Learning from Human Feedbacktechniqueactive
Reinforcement Learning with Verifiable Rewardstechniqueactive
Responsible Scaling Policyconceptactive
Retrieval-Augmented Generationtechniqueactive
Rotary Position Embeddingtechniqueactive
Sakana: Fugu Maxmodelactive
Sakana: Fugu Ultra v2modelactive
Scaling lawsconceptactive
SGLangtoolactive
Space Bunny Alphamodelactive
SpaceXAIorgactive
SpaceXAI: Grok 4.6modelactive
SpaceXAI: Grok 4.7modelactive
Speculative decodingtechniqueactive
Structured generationtechniqueactive
Temperature and top-pconceptactive
Tencentorgactive
Tencent: Hy4 previewmodelactive
Test-time scalingtechniqueactive
The ALPAC Reporteventdead
The Bitter Lessonconceptactive
The ChatGPT Launcheventdead
The GPT-2 Staged Releaseeventdead
The Hugging Face Agent Intrusioneventdead
The Lighthill Reporteventdead
The reversal curseconceptactive
The Stable Diffusion Releaseeventdead
Thinking Machines Laborgactive
Thinking Machines: Inkling (batch)modelactive
Thinking Machines: Inkling Small (batch)modelactive
Tokenizationconceptactive
Upstage: Solar Mini 4modelactive
vLLMtoolactive
Xiaomi: MiMo-V2.6-Flashmodelactive
Xiaomi: MiMo-V2.6-Promodelactive
Xiaomi: MiMo-V2.6-Pro-UltraSpeedmodelactive
Z.aiorgactive
Z.ai: GLM 4.5modeldeprecated
Z.ai: GLM 4.5Vmodelactive
Z.ai: GLM 4.7 Flashmodelactive
Z.ai: GLM 5.1modelactive
Z.ai: GLM 5.2 (batch)modelactive
Z.ai: GLM 5.2 (free)modelactive
Z.ai: GLM 5.3 (batch)modelactive
Z.ai: GLM 5.3 FlashXmodelactive
Z.ai: GLM 5.3 Primemodelactive

397 further entries are data-only stubs — minted mechanically from the source registry, not listed here, and reachable through the search box above and the open dataset. An entry is listed once it has a prose body, or two facts and a timeline event, or a lifecycle status worth keeping.