Artificial Analysis Agentic Index (Maximum Reasoning)

33 models

	Rank	Model	Score
🇺🇸	1	Anthropic: Claude Fable 5	80.6
🇺🇸	2	Anthropic: Claude Opus 4.8	77.8
🇺🇸	3	OpenAI: GPT-5.5	74.1
🇺🇸	4	Anthropic: Claude Opus 4.7	71.3
🇺🇸	5	Google: Gemini 3.5 Flash	70.3
🇺🇸	6	OpenAI: GPT-5.4	68
🇨🇳	7	Xiaomi: MiMo-V2.5 Pro小米集团	67.4
🇨🇳	8	DeepSeek: DeepSeek V4 Pro深度求索	67.2
🇨🇳	9	Z.ai: GLM 5.1北京智谱华章科技有限公司	67.1
🇨🇳	10	Qwen: Qwen3.7 Max通义千问	66.6
🇨🇳	11	Moonshot AI: Kimi K2.6月之暗面	66
🇺🇸	12	SpaceXAI: Grok 4.3	65.9
🇨🇳	13	Qwen: qwen3.6-max通义千问	64.8
🇺🇸	14	Anthropic: Claude Sonnet 4.6	63
🇺🇸	15	Meta: Muse Spark	62
🇨🇳	16	MiniMax: minimax-m2.7稀宇科技	61.5
🇺🇸	17	OpenAI: gpt-5.3-codex	60.5
🇺🇸	18	OpenAI: gpt-5.2	60.2
🇺🇸	19	OpenAI: gpt-5.2-codex	56.5
🇫🇷	20	Mistral AI: mistral-medium-3.5-128b	53.2
🇺🇸	21	SpaceXAI: Grok 4.1 Fast	49.3
🇺🇸	22	SpaceXAI: Grok 4 Fast	39.5
🇺🇸	23	SpaceXAI: grok-code-fast-1	35.6
🇺🇸	24	Google: Gemma 4 26B A4B Instruct	32.1
🇺🇸	25	OpenAI: o1	31.1
❓	26	NousResearch: hermes-4-405b	12.6
❓	27	NousResearch: hermes-4-70b	11.7
🇦🇪	28	Technology Innovation Institute: falcon-h1r-7b	9.3
🇺🇸	29	Meta Llama: Llama 3.3 70B Instruct	9.1
🇺🇸	30	Meta Llama: Llama 4 Maverick Instruct	7.2
🇨🇳	31	Qwen: Qwen3 0.6B通义千问	7
🇺🇸	32	Google: Gemma 4 E2B Instruct	6.9
🇨🇳	33	DeepSeek: DeepSeek R1深度求索	3.8