Benchmark · general

Artificial Analysis Intelligence Index

16 results 14 models

Artificial Analysis's independent composite index averaging ~10 evals into one 0-100 index score in points (frontier models ~50-67), not a percentage. Distinct from its sibling AA Coding Agent / Agentic indices.

DeepSeek V4-Flash Claude Opus 5 Grok 4.6 Solar Pro 4 Gemini 3.7 Flash Muse Spark 1.2 Grok 4.5 Qwen3.8-Max Muse Spark 1.1 Qwen 3.8 27B GLM-5.3-Flash GLM-5.3 GPT-6 Astra Claude Fable 5.1

Timeline

Date	Model	Score	Source
2026-09-03	GPT-6 Astra	61.0%	OpenAI releases GPT-6 Astra with ARC-AGI 3 and security improvements
2026-09-03	Claude Fable 5.1	66.0%	OpenAI releases GPT-6 Astra with ARC-AGI 3 and security improvements
2026-08-28	GLM-5.3	60.0%	Z.ai fine-tunes GLM-5.3 for cybersecurity gains
2026-08-27	GLM-5.3-Flash	57.0%	NVIDIA buys HuggingFace for $13B; Z.ai launches GLM-5.3-Flash
2026-08-27	GLM-5.3-Flash	57.0%	Z.ai launches GLM-5.3-Flash, a 320B-parameter multimodal model with 1M context
2026-08-21	Grok 4.6	61.0%	SpaceXAI introduces Grok 4.6 with Cursor data for agentic work
2026-08-18	Qwen 3.8 27B	52.0%	Qwen 3.8 27B scores 52 on the Artificial Analysis Intelligence Index
2026-08-14	Muse Spark 1.2	57.0%	Meta releases Muse Code agentic harness and low-cost Muse Spark 1.2 model
2026-08-14	Grok 4.5	56.0%	Meta releases Muse Code agentic harness and low-cost Muse Spark 1.2 model
2026-08-14	Qwen3.8-Max	58.0%	Meta releases Muse Code agentic harness and low-cost Muse Spark 1.2 model
2026-08-14	Muse Spark 1.1	53.0%	Meta releases Muse Code agentic harness and low-cost Muse Spark 1.2 model
2026-08-14	Gemini 3.7 Flash	56.0%	Google releases Gemini 3.7 Flash with algorithmic improvements and lower pricing
2026-08-13	Grok 4.6	61.0%	xAI releases Grok 4.6; Alibaba opens Qwen3.8-MoE; DeepSeek V4 Pro GA
2026-08-13	Solar Pro 4	42.0%	xAI releases Grok 4.6; Alibaba opens Qwen3.8-MoE; DeepSeek V4 Pro GA
2026-07-31	Claude Opus 5	61.0%	Anthropic launches Claude Opus 5; OpenAI models breach Hugging Face during security test
2026-07-31	DeepSeek V4-Flash	50.0%	DeepSeek V4-Flash achieves 50 on ArtificialAnalysis Index