Explore and compare LLMs

Q: How do I compare LLMs?

Sort by Quality Index to find the strongest models overall, then filter by price, speed, or context window to match your constraints. Use the Compare page to put 2–4 finalists head to head.

Find the models that fit your workload by capability, task economics, speed, context, and provider availability.

Artificial Analysis · snapshot synced Aug 27, 2026, 20:09 UTCAA data · Aug 27, 20:09 UTC185 models · 87 providersLatest tracked Aug 26, 2026 · Index v4.1

Rank by

Show

Model rankings

114 results · sorted by intelligence

Open a row for metrics and endpoints

#
1	Kimi K3OpenKimi	59.7	76.2	54.3	$0.838	71s	1.0M	11
2	GLM-5.3OpenZ AI	59.5	74.8	59.1	$0.683	30s	1.0M	1
3	Qwen3.8 MaxOpenAlibaba	58.1	71.8	58.4	$0.913	2.1 min	1.0M	1
4	Qwen3.8 2.4T A95BOpenAlibaba	57.7	71.9	57.1	$0.807	1.9 min	984K	6
5	GLM-5.3-FlashOpenZ AI	57.5	71.5	58.2	$0.087	52s	1.0M	11
6	DeepSeek V4 Pro 0813OpenDeepSeek	53.2	68.8	49.6	$0.265	38s	1.0M	8
7	GLM-5.2OpenZ AI	52.6	68.8	45.7	$0.445	6.7s	1.0M	20
8	Qwen3.8 27B (xhigh)OpenAlibaba	52	68.1	50.9	$0.369	51s	1.0M	2
9	DeepSeek V4 Flash 0731OpenDeepSeek	51.8	69.1	48.4	$0.112	22s	1.0M	14
10	DeepSeek V4 Flash VisionOpenDeepSeek	51.5	65	52.9	$0.116	22s	1.0M	1
11	Kimi K3 (low)OpenKimi	48.3	72	39.6	$0.242	74s	1.0M	2
12	MiniMax-M3OpenMiniMax	45.4	58.6	36.1	$0.139	22s	1.0M	10
13	DeepSeek V4 ProOpenDeepSeek	45.3	59.4	37.8	$0.043	8.8s	1.0M	9
14	Qwen3.8 27B (medium)OpenAlibaba	44.5	56.1	49.8	$0.284	50s	1.0M	1
15	Kimi K2.7 CodeOpenKimi	43	60.8	30.3	$0.222	64s	262K	10
16	MiMo-V2.5-ProOpenXiaomi	42.9	60.2	29.5	$0.034	19s	1.0M	3
17	Qwen3.8 27B (low)OpenAlibaba	42.9	58.2	43.7	$0.265	48s	1.0M	1
18	Inkling (xhigh)OpenThinking Machines	42.3	52.1	34.1	$0.339	52s	524K	4
19	Hy3OpenTencent	42.2	58.8	31.4	$0.036	39s	262K	4
20	Nex-N2-ProOpenNex AGI	41.7	59.1	31.2	N/A	21s	262K	1
21	Inkling SmallOpenThinking Machines	41.2	52.9	31.9	$0.073	22s	524K	3
22	Agnes 2.5 Pro AlphaOpenSapiens AI	39.7	58.8	25.4	N/A	21s	1.0M	1
23	Qwen3.7 PlusOpenAlibaba	39.4	55.9	20.7	$0.170	47s	1.0M	1
24	Nemotron 3 Ultra 550B A55B (Reasoning)OpenNVIDIA	38.3	49.3	27.5	$0.383	18s	512K	4
25	MiMo-V2.5OpenXiaomi	38	56.8	24.4	$0.010	39s	1.0M	4
26	Ling 3.0 FlashOpenInclusionAI	37.8	50.6	29.3	$0.038	8.9s	262K	2
27	Qwen3.6 27B (Reasoning)OpenAlibaba	37.7	53.7	27.5	$0.292	2.1 min	262K	6
28	Muse Glimmer (high)OpenMeta	35.1	49	22.9	$0.060	24s	131K	2
29	Qwen3.8 27BOpenAlibaba	34.7	44.6	30.4	$0.426	12s	1.0M	1
30	Qwen3.5 397B A17B (Reasoning)OpenAlibaba	34.3	48.2	19.8	$0.357	42s	262K	8
31	LongCat 2.0OpenLongCat	34	45.3	22	$0.121	63s	1.0M	1
32	Qwen3.5 122B A10B (Reasoning)OpenAlibaba	32.8	45.7	21.3	$0.253	21s	262K	4
33	Qwen3.5 397B A17BOpenAlibaba	32.7	N/A	N/A	N/A	8.2s	262K	6
34	Qwen3.6 35B A3B (Reasoning)OpenAlibaba	32.1	41.9	21.6	$0.294	51s	262K	6
35	Ring-2.6-1TOpenInclusionAI	31.7	42.8	20.1	$0.360	24s	262K	1
36	Qwen3.5 Omni PlusOpenAlibaba	31.3	N/A	N/A	N/A	6.1s	256K	1
37	Qwen3.6 27BOpenAlibaba	31.3	46.6	23.6	$0.398	13s	262K	5
38	Step 3.7 FlashOpenStepFun	30.9	39.6	21.7	$0.091	30s	256K	2
39	Mistral Medium 3.5OpenMistral	30.4	46.9	19.2	$0.406	20s	262K	1
40	Gemma 4 31B (Reasoning)OpenGoogle	29.7	43.4	14.4	$0.0000	64s	262K	12
41	DeepSeek V4 FlashOpenDeepSeek	29.3	N/A	N/A	N/A	N/A	1.0M	2
42	Qwen3.5 122B A10BOpenAlibaba	28.2	43.3	16.2	$0.195	5.8s	262K	2
43	Gemma 4 26B A4B (Reasoning)OpenGoogle	26.1	39.3	11	$0.039	N/A	1.0M	5
44	Nemotron 3 Super 120B A12B (Reasoning)OpenNVIDIA	25.7	37.7	8.8	$0.229	19s	262K	3
45	Qwen3.6 35B A3BOpenAlibaba	24.6	28.1	19.2	$0.630	5.6s	262K	5
46	Ling 3.0 TinyOpenInclusionAI	24.5	26.5	16	$0.0000	19s	262K	1
47	Qwen3.5 35B A3BOpenAlibaba	24.3	37	11.8	$0.241	5.0s	262K	2
48	gpt-oss-120b (high)OpenOpenAI	24.1	30.4	13.4	$0.073	16s	131K	14
49	Nemotron 3.5 LightningOpenNVIDIA	23.6	26.8	13.8	$0.077	9.3s	1.0M	5
50	Command A+OpenCohere	22.8	27.8	9.2	$0.0000	10s	200K	1
51	Gemma 4 31BOpenGoogle	22.3	33.2	11.1	$0.044	7.0s	262K	9
52	Gemma 4 12B (Reasoning)OpenGoogle	22.2	31	N/A	N/A	25s	262K	1
53	Qwen3.5 9B (Reasoning)OpenAlibaba	21.8	28.7	7	$0.242	35s	262K	2
54	Qwen3 Coder NextOpenAlibaba	21.3	36.2	8.9	$0.337	5.7s	262K	3
55	Qwen3.5 9BOpenAlibaba	20.6	23.5	N/A	N/A	6.1s	262K	1
56	Qwen3.5 4B (Reasoning)OpenAlibaba	20.4	22.6	N/A	N/A	59s	262K	1
57	Gemma 4 26B A4BOpenGoogle	20.4	N/A	N/A	N/A	8.3s	1.0M	6
58	North Mini CodeOpenCohere	20.2	36.5	3.1	$0.0000	83s	256K	1
59	Mistral Small 4 (Reasoning)OpenMistral	19.7	26.6	4.6	$0.100	16s	256K	1
60	Granite 4.2 8BOpenIBM	19.6	22.4	9.2	$0.016	23s	131K	2
61	Devstral 2OpenMistral	19.2	31.3	10.6	$0.0000	11s	262K	1
62	Qwen3.5 Omni FlashOpenAlibaba	19.2	N/A	N/A	N/A	3.8s	256K	1
63	Trinity Large ThinkingOpenArcee AI	18.7	25.8	3.7	$0.164	16s	262K	2
64	HyperNova 60B 2605OpenMultiverse Computing	18.3	23.2	6.8	$0.024	7.7s	131K	1
65	Devstral Small 2OpenMistral	17.7	29.3	10.8	$0.0000	5.9s	256K	1
66	Qwen3 Next 80B A3B (Reasoning)OpenAlibaba	16.9	17.4	2.1	$0.042	15s	262K	3
67	Qwen3.5 4BOpenAlibaba	16.1	20.3	N/A	N/A	18s	262K	1
68	Mistral Large 3OpenMistral	15.9	20.1	5.5	$0.081	14s	262K	3
69	gpt-oss-20b (high)OpenOpenAI	15.2	20.7	3.1	$0.025	22s	131K	6
70	Nemotron 3 Nano Omni 30B A3B ReasoningOpenNVIDIA	15	13.8	N/A	N/A	8.6s	66K	1
71	gpt-oss-120b (low)OpenOpenAI	14.9	21.2	1	$0.020	16s	131K	13
72	NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)OpenNVIDIA	14.5	14.4	2	$0.025	12s	262K	3
73	Llama 4 MaverickOpenMeta	14.5	16.3	1.2	$0.035	5.9s	1.0M	5
74	gpt-oss-20b (low)OpenOpenAI	14.4	N/A	N/A	N/A	18s	131K	5
75	Granite 4.2 3BOpenIBM	14.3	17.5	1.9	$0.0047	13s	131K	1
76	Qwen3 Next 80B A3B InstructOpenAlibaba	13.8	N/A	N/A	N/A	5.0s	262K	6
77	Gemma 4 12BOpenGoogle	13.2	N/A	N/A	N/A	6.6s	262K	1
78	Llama Nemotron Super 49B v1.5 (Reasoning)OpenNVIDIA	12.4	N/A	N/A	N/A	33s	131K	1
79	Mistral Small 4OpenMistral	12.3	N/A	N/A	N/A	4.1s	262K	1
80	Gemma 4 E4B (Reasoning)OpenGoogle	12.2	9.4	N/A	N/A	52s	262K	1
81	Magistral Small 1.2OpenMistral	11.5	14.7	1.5	$0.295	N/A	128K	1
82	Ministral 3 14BOpenMistral	11.2	14.4	2.2	$0.158	6.2s	262K	2
83	Llama 4 ScoutOpenMeta	10.3	8.2	1.1	$0.011	4.5s	1.3M	6
84	Hermes 4 - Llama-3.1 70B (Reasoning)OpenNous Research	9.9	N/A	N/A	N/A	33s	128K	1
85	Qwen3 Omni 30B A3B (Reasoning)OpenAlibaba	9.5	N/A	N/A	N/A	26s	66K	1
86	Llama 3.3 Instruct 70BOpenMeta	9.3	11.9	N/A	N/A	7.8s	131K	12
87	Ministral 3 8BOpenMistral	9	9.7	1.2	$0.185	4.8s	262K	2
88	Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)OpenNVIDIA	8.9	N/A	N/A	N/A	50s	131K	1
89	NVIDIA Nemotron Nano 12B v2 VL (Reasoning)OpenNVIDIA	8.8	N/A	N/A	N/A	34s	131K	1
90	Hermes 4 - Llama-3.1 405B (Reasoning)OpenNous Research	8.8	N/A	N/A	N/A	75s	128K	1
91	Gemma 4 E4BOpenGoogle	8.7	N/A	N/A	N/A	10s	262K	1
92	NVIDIA Nemotron Nano 9B V2 (Reasoning)OpenNVIDIA	8.7	N/A	N/A	N/A	14s	131K	1
93	Hermes 4 - Llama-3.1 405BOpenNous Research	8.6	N/A	N/A	N/A	18s	128K	1
94	Llama Nemotron Super 49B v1.5OpenNVIDIA	8.5	N/A	N/A	N/A	8.3s	131K	1
95	LFM2.5-8B-A1BOpenLiquid AI	8.1	N/A	N/A	N/A	8.8s	33K	1
96	Command AOpenCohere	7.5	N/A	N/A	N/A	9.1s	288K	2
97	Llama 3.1 Nemotron Instruct 70BOpenNVIDIA	7.4	N/A	N/A	N/A	12s	131K	1
98	NVIDIA Nemotron 3 Nano 30B A3BOpenNVIDIA	7.2	N/A	N/A	N/A	3.2s	262K	2
99	NVIDIA Nemotron Nano 9B V2OpenNVIDIA	7.2	N/A	N/A	N/A	4.2s	131K	2
100	Ministral 3 3BOpenMistral	7.1	4.8	1.6	$0.130	3.2s	256K	2
101	Hermes 4 - Llama-3.1 70BOpenNous Research	6.7	N/A	N/A	N/A	7.4s	128K	1
102	Granite 4.1 8BOpenIBM	6.4	9.5	N/A	N/A	4.9s	131K	1
103	Phi-4 Mini InstructOpenMicrosoft	5.7	3.8	N/A	N/A	12s	128K	1
104	Granite 4.0 H SmallOpenIBM	4.9	N/A	N/A	N/A	44s	128K	1
105	Qwen3 Omni 30B A3B InstructOpenAlibaba	4.8	N/A	N/A	N/A	7.3s	66K	1
106	Phi-4OpenMicrosoft	4.6	N/A	N/A	N/A	15s	16K	2
107	Phi-4 Multimodal InstructOpenMicrosoft	4.2	N/A	N/A	N/A	29s	128K	1
108	NVIDIA Nemotron Nano 12B v2 VLOpenNVIDIA	4.2	N/A	N/A	N/A	4.8s	131K	2
109	Reka Flash 3OpenReka AI	3.7	N/A	N/A	N/A	N/A	54K	1
110	Llama 3.2 Instruct 11B (Vision)OpenMeta	3	N/A	N/A	N/A	11s	131K	1
111	LFM2.5-VL-1.6BOpenLiquid AI	1	N/A	N/A	N/A	2.6s	33K	1
112	Tiny Aya GlobalOpenCohere	1	N/A	N/A	N/A	N/A	8K	1
113	Granite 4.2 30BOpenIBM	0	N/A	N/A	N/A	N/A	131K	1
114	Qwen3.8-Flash-NextOpenAlibaba	0	N/A	N/A	N/A	N/A	1.0M	1

How to use this LLM ranking page

Start with capability, then force the economics

Frontier models often cluster tightly on raw intelligence. Cost per task, response time, provider coverage, and context length usually create the real shortlist.

Move from discovery to a defensible final pick

Use Explore to find the shape of the market, then move into Compare or use the Agentic Fit Finder when task-level reliability and cost matter more than chat quality alone.

Frequently Asked Questions

What is the best LLM in 2026?

Claude Opus 5 leads the overall quality ranking right now. The best model for you depends on your use case — coding, cost, speed, and context length all shift the answer.

How do I compare LLMs?

Sort by Quality Index for overall strength, then filter by price, speed, or context window to match your constraints. Move to the Compare page to put 2–4 finalists head to head.

Which AI model has the best quality-to-price ratio?

Open-weight models like DeepSeek, Qwen, and Llama often lead on quality-to-price. For agentic workflows, cost per task is usually a better filter than token price alone.

How often is this leaderboard updated?

Data is pulled from Artificial Analysis and refreshed automatically. New models appear as soon as they have benchmark scores and provider endpoints.

Model rankings

Current live rankings

Browse the latest ranking pages for overall models, coding, local hardware, open source, Ollama, long context, and agentic workflows.

Coding

Best LLM for Coding

Current coding leaderboard using LiveCodeBench, Terminal-Bench, and SciCode.

Open page →

Open Source

Best Open Source LLM

Top open-weight models for self-hosting, Ollama, and low-cost API use.

Open page →

Self-Hosted

Best Local LLM

Best local AI models by hardware tier for self-hosting on Macs, RTX GPUs, and workstations.

Open page →

Local Coding

Best Local LLM for Coding

Coding-focused local models mapped to 8GB, 24GB, 64GB, and server-class hardware.

Open page →

Ollama

Best Ollama Models

Ollama-first picks for coding, chat, reasoning, and low-friction local inference.

Open page →

Long Context

Largest Context Window LLM

Best long-context models for large documents, codebases, and retrieval-heavy workflows.

Open page →

Agents

Best Agentic Models

Rankings for tool use, multi-step execution, and autonomous agent workflows.

Open page →