⌕Is ___ SOTA?
Categories · Oct 14, 2026

What can be state of the art

Each category has its own evidence and its own crown. Open a category to change the slice: overall, open weights, ≤32B, or best per dollar.

General LLM
General LLM
Gemini 4 Argon
12d reign2026-10-02
Reasoning / knowledge
ReasoningCO-SOTA
Claude Opus 5.5 & Gemini 4 Argon
12d reign2026-10-02
MathCO-SOTA
Gemini 3.1 Pro Preview & GPT-5.4 Pro & Muse Spark
181d reign2026-04-16
Science
No crown yet
Waiting on a primary board
Coding
CodingCO-SOTA
Claude Opus 5.5 & Claude Sonnet 5.5
8d reign2026-10-06
Coding › AgenticCO-SOTA
Claude Opus 5.5 & Claude Sonnet 5.5
8d reign2026-10-06
Coding › Web dev
Claude Opus 5.5
15d reign2026-09-29
Long context
Long contextCO-SOTA
Claude Opus 5.5 & Gemini 4 Argon
12d reign2026-10-02
Vision
VisionCO-SOTA
GPT-5.6 Sol & GPT-6 Astra
35d reign2026-09-09
Image generation
Image generation
GPT Image 2.5 Sunburst
20d reign2026-09-24
Image edit
GPT Image 2.5 Sunburst
22d reign2026-09-22
Video generation
Video generation
Veo 3
433d reign2025-08-07
Factuality
FactualityCO-SOTA
Claude Fable 5.1 & Claude Opus 5.5 & Gemini 4 Argon
12d reign2026-10-02
Abstraction
Abstraction (ARC)
No crown yet
Waiting on a primary board
Embeddings / retrieval
EmbeddingsCO-SOTA
Microsoft Harrier Oss V1 27B & Qwen Qwen3 Embedding 8B & Tencent Kalm Embedding Gemma3 12B 2511
196d reign2026-04-01
Speech
Speech-to-text
No crown yet
Waiting on a primary board