AA-LCR leaderboard

A display-only Artificial Analysis long-context reasoning evaluation.

Models ranked by their published Artificial Analysis Long Context Reasoning value

Rank	Model	Provider	Accuracy
1	Kimi K3	Moonshot AI	88.7
2	Claude Fable 5.1	Anthropic	85.3
3	GPT-5.5	OpenAI	84.3
4	DeepSeek V4.1 Flash	DeepSeek	84.0
4	GPT-5.6 Sol	OpenAI	84.0
6	GPT-5.6 Luna	OpenAI	83.7
7	GPT-5.3 Codex	OpenAI	83.3
7	Muse Glimmer 30B	Meta	83.3
9	GPT-5.6 Terra	OpenAI	83.0
9	MiniMax M3	MiniMax	83.0
9	Muse Spark 1.3	Meta	83.0
12	GPT-5.2	OpenAI	82.7
13	Claude Fable 5	Anthropic	82.3
13	GPT-5.2-Codex	OpenAI	82.3
15	Claude Sonnet 5	Anthropic	82.0
15	Gemini 3.1 Pro	Google	82.0
15	GPT-5.4	OpenAI	82.0
15	Qwen3.8-27B	Alibaba	82.0
19	Gemini 3.7 Flash	Google	81.7
20	Gemini 3.8 Flash	Google	81.3
21	Kimi K2.6	Moonshot AI	81.0
22	GPT-6 Astra	OpenAI	80.7
22	Qwen 3.6 Max (preview)	Alibaba	80.7
24	DeepSeek V4 Pro 0813	DeepSeek	80.3
24	Grok 4.6	xAI	80.3
24	Qwen3.8 Max Preview	Alibaba	80.3
27	Gemini 3.6 Flash	Google	80.0
27	GPT-5.1	OpenAI	80.0
29	DeepSeek V4 Flash 0731	DeepSeek	79.7
29	GLM-5.3	Z.AI	79.7
29	MiMo-V2.5-Pro	Xiaomi	79.7
29	Qwen3.8-Flash-Next	Alibaba	79.7
33	APApodex 1.1	Apodex	79.3
33	APApodex 1.1 Mini	Apodex	79.3
33	Claude Opus 5	Anthropic	79.3
33	Grok 4.5	xAI	79.3
33	Kimi K2.7 Code	Moonshot AI	79.3
38	Hy3	Tencent	79.0
38	Muse Spark 1.2	Meta	79.0
38	Qwen3.7 Max	Alibaba	79.0
41	Claude Opus 4.7 (Adaptive)	Anthropic	78.7
42	GLM-5.2	Z.AI	78.3
42	Ling 3.0 Flash VL	InclusionAI	78.3
42	MiniMax M2.7	MiniMax	78.3
42	Qwen3.6 Plus	Alibaba	78.3
46	GPT-5 (high)	OpenAI	78.2
47	Claude Opus 4.6 (Adaptive)	Anthropic	78.0
47	Kimi K2.5	Moonshot AI	78.0
47	Kimi K2.5 (Reasoning)	Moonshot AI	78.0
47	Muse Spark	Meta	78.0
51	Claude Opus 4.8	Anthropic	77.7
51	Muse Spark 1.1	Meta	77.7
51	Qwen3.5-27B	Alibaba	77.7
54	Claude Opus 4.5 Thinking	Anthropic	77.3
54	TMInkling	Thinking Machines Lab	77.3
54	Qwen3.6-27B	Alibaba	77.3
57	GPT-5.4 mini	OpenAI	77.0
57	PMTernary Bonsai 2 27B	Prism ML	77.0
59	GPT-5.4 nano	OpenAI	76.7
60	MCQuasar 438B	Multiverse Computing	76.3
60	Qwen3.5-122B-A10B	Alibaba	76.3
62	Claude 4.1 Opus Thinking	Anthropic	76.0
62	Gemini 3 Pro	Google	76.0
62	Gemini 3.5 Flash-Lite	Google	76.0
62	GPT-5 (medium)	OpenAI	76.0
66	Claude Opus 4.7	Anthropic	75.7
66	GLM-5	Z.AI	75.7
66	TMInkling-Small	Thinking Machines Lab	75.7
69	MiMo-V2-Omni	Xiaomi	75.0
70	o3	OpenAI	74.7
71	Grok 4.1 Fast (Reasoning)	xAI	74.0
72	GLM-5.1	Z.AI	73.7
72	Grok 4 Fast (Reasoning)	xAI	73.7
72	Step 3.7 Flash	StepFun	73.7
75	Ling 3.0 Flash	InclusionAI	73.0
75	Ling 3.0 Flash FP8	InclusionAI	73.0
75	Qwen3.7 Plus	Alibaba	73.0
78	Qwen3.5-35B-A3B	Alibaba	72.0
79	GLM-5-Turbo	Z.AI	71.7
79	Qwen3.6-35B-A3B	Alibaba	71.7
81	GLM-4.7	Z.AI	71.0
81	Solar Pro 4	Upstage	71.0
83	Claude Opus 4.5	Anthropic	70.7
84	GLM-5V-Turbo	Z.AI	70.3
85	Gemma 4 31B	Google	69.7
86	Gemini 3.5 Flash	Google	69.3
86	GPT-5.1-Codex	OpenAI	69.3
86	GPT-5.1-Codex-Max	OpenAI	69.3
86	Mistral Medium 3.5 128B	Mistral	69.3
90	Gemini 2.5 Pro	Google	69.0
91	Claude Sonnet 4.6	Anthropic	68.3
91	GPT-4.1	OpenAI	68.3
91	MiMo-V2-Pro	Xiaomi	68.3
94	Grok 4	xAI	68.0
94	Mercury 2.5	Inception	68.0
96	Claude Opus 4.6	Anthropic	67.0
96	Nemotron 3 Ultra	NVIDIA	67.0
98	Hy3 Preview	Tencent	66.7
99	STA.X K2	SK Telecom	66.0
100	Gemma 4 26B A4B	Google	65.7
100	Nemotron 3 Super 100B	NVIDIA	65.7
102	o1	OpenAI	65.0
103	Grok 4.3	xAI	64.3
103	Qwen3.5 397B	Alibaba	64.3
103	Qwen3.5 397B (Reasoning)	Alibaba	64.3
106	Gemma 4 12B	Google	63.7
107	Solar Open 2	Upstage	62.3
108	K-Exaone	LG AI Research	61.3
109	Ling 3.0 Tiny	InclusionAI	60.3
110	OPMiniCPM5-2B	OpenBMB	59.0
111	DeepSeek V3.1 (Reasoning)	DeepSeek	56.7
112	K-EXAONE 2.0	LG AI Research	56.2
113	DeepSeek-R1	DeepSeek	55.7
114	Gemini 3 Flash	Google	55.3
115	Grok Code Fast 1	xAI	53.0
115	Kimi K2	Moonshot AI	53.0
117	Command A+	Cohere	52.7
118	GPT-OSS 120B	OpenAI	52.0
119	Llama 4 Maverick	Meta	50.0
119	Qwen3 Max	Alibaba	50.0
121	Gemini 2.5 Flash	Google	49.9
122	Mistral Small 4	Mistral	49.7
122	Mistral Small 4 (Reasoning)	Mistral	49.7
124	GPT-4o	OpenAI	49.3
125	Nemotron 3.5 Lightning 30B A3B NVFP4	NVIDIA	49.2
126	Granite 4.2 30B	IBM	49.0
127	DeepSeek V3.1	DeepSeek	47.0
128	GLM-4.5-Air	Z.AI	46.7
129	DeepSeek V3.2	DeepSeek	45.7
130	Granite 4.2 8B	IBM	45.0
131	Claude 4 Sonnet	Anthropic	44.0
131	GPT-4.1 mini	OpenAI	44.0
133	Nemotron 3 Nano Omni 30B A3B	NVIDIA	39.7
134	CECeleris-1	Celeris	38.0
134	Nemotron 3 Nano 30B	NVIDIA	38.0
134	Trinity-Large-Preview	Arcee AI	38.0
134	Trinity-Large-Thinking	Arcee AI	38.0
138	North Mini Code	Cohere	37.3
139	MiMo-V2-Flash	Xiaomi	36.3
140	Mistral Large 3	Mistral	36.0
141	GPT-OSS 20B	OpenAI	34.7
142	Solar Pro 3	Upstage	32.3
143	Gemma 4 E4B	Google	32.0
144	Grok 4.1 Fast	xAI	31.3
144	Ling 2.6 Flash	InclusionAI	31.3
144	Mistral Medium 3	Mistral	31.3
147	DeepSeek V3	DeepSeek	29.3
148	Claude 3 Haiku	Anthropic	27.7
148	Llama 4 Scout	Meta	27.7
150	GLM-4.6	Z.AI	26.3
151	Llama 3.1 405B	Meta	25.3
152	Granite 4.2 3B	IBM	24.3
153	Nova Pro	Amazon	21.0
154	GPT-4.1 nano	OpenAI	20.3
155	Gemma 4 E2B	Google	16.3
156	FAUltravox v0.6 Llama 3.3 70B	Fixie AI	15.7
157	DeepSeek R1 Distill Qwen 32B	DeepSeek	8.7
158	Gemma 3 27B	Google	7.3
159	Granite-4.0-H-1B	IBM	7.0
160	Granite-4.0-1B	IBM	6.0
161	LFM2.5-2.6B	LiquidAI	5.7
162	Exaone 4.0 1.2B	LG AI Research	0.0
162	Granite-4.0-350M	IBM	0.0
162	Granite-4.0-H-350M	IBM	0.0
162	LFM2.5-8B-A1B	LiquidAI	0.0
162	LFM2.5-VL-1.6B-Extract	LiquidAI	0.0
162	Phi-4	Microsoft	0.0
162	Qwen3-Omni-30B-A3B-Instruct	Alibaba	0.0
162	Qwen3-Omni-30B-A3B-Thinking	Alibaba	0.0
162	SASarvam 105B	Sarvam	0.0
162	SASarvam 30B	Sarvam	0.0
162	Solar Pro 2	Upstage	0.0

Evidence key: ObservedLast good

Rows are ordered by the value Artificial Analysis model benchmarks published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.