AI Value Index — Best AI Model Rankings & Benchmark Leaderboard

Find the best AI model for YOUR use case. Pick your benchmarks, set your weights, see actual data — not abstract scores.

Metrics & Weights

6 metrics selected — total weight: 100%

Preset

Active metrics

SWE-bench Verified 30%HumanEval+ 15%Chatbot Arena ELO 15%Output Speed 15%Input Cost 13%Output Cost 13%

#	Model		30%	15%	15%	15%	12.5%	12.5%
1	Gemini 3 Flash Google	82.4	78.0%	86.0%	1473	200	$0.50	$3.0
2	GPT-5.2 OpenAI	80.2	80.0%	95.0%	1475	90	$1.8	$14.0
3	Gemini 3.1 Pro Google	79.7	80.6%	93.0%	1501	65	$2.0	$12.0
4	GPT-5.1 OpenAI	78.4	76.3%	93.0%	1464	95	$1.3	$10.0
5	Claude Opus 4.6 Anthropic	78.3	80.8%	93.5%	1496	68	$5.0	$25.0
6	GPT-5.1 Codex OpenAI	77	78.0%	96.0%	1395	85	$1.3	$10.0
7	Claude Opus 4.5 Anthropic	75.8	80.9%	92.0%	1468	50	$5.0	$25.0
8	GPT-5 OpenAI	75	74.9%	92.0%	1390	100	$1.3	$10.0
9	Claude Sonnet 4.6 Anthropic	73.7	79.6%	91.0%	1395	57	$3.0	$15.0
10	Grok 4.1 Fast xAI	73.7	60.0%	88.0%	1482	120	$0.20	$0.50
11	Gemini 3 Pro Google	72.8	65.0%	91.0%	1492	60	$2.0	$12.0
12	Claude Sonnet 4.5 Anthropic	72.4	77.2%	90.0%	1380	67	$3.0	$15.0
13	Grok 4 xAI	71.9	72.0%	90.0%	1430	55	$3.0	$15.0
14	DeepSeek V3.2 DeepSeek	71.3	73.0%	87.0%	1370	70	$0.14	$0.28
15	o4 Mini OpenAI	71	68.1%	88.0%	1350	120	$1.1	$4.4
16	Gemini 2.5 Pro Google	70.7	63.8%	89.0%	1465	55	$1.3	$10.0
17	Qwen 3.5 397B Qwen	70.6	76.4%	88.0%	1350	45	$0.60	$3.6
18	Claude Sonnet 4 Anthropic	69.9	72.7%	88.0%	1365	75	$3.0	$15.0
19	o3 OpenAI	69	69.1%	90.0%	1380	40	$2.0	$8.0
20	GPT-5.1 Codex Mini OpenAI	67.5	55.0%	88.0%	1310	170	$0.25	$2.0
21	Qwen 3 Coder Qwen	67.4	66.5%	90.0%	1290	80	$0.20	$0.80
22	Claude 3.7 Sonnet Anthropic	64.3	62.3%	86.0%	1340	70	$3.0	$15.0
23	GPT-5 Mini OpenAI	64	48.0%	85.0%	1300	180	$0.25	$2.0
24	DeepSeek R1 0528 DeepSeek	63.9	55.0%	88.0%	1375	40	$0.55	$2.2
25	Grok 4 Fast xAI	63.6	48.0%	85.0%	1370	110	$0.20	$0.50
26	Gemini 2.5 Flash Google	63.5	38.0%	82.0%	1320	251	$0.30	$2.5
27	Claude Opus 4 Anthropic	62.7	72.5%	90.0%	1375	50	$15.0	$75.0
28	o3 Mini OpenAI	62.1	50.0%	87.0%	1320	100	$1.1	$4.4
29	GPT-4.1 OpenAI	61.8	50.0%	89.0%	1340	70	$2.0	$8.0
30	o3 Pro OpenAI	61.5	70.0%	92.0%	1410	25	$20.0	$80.0
31	Mistral Large 25.12 Mistral	60.2	42.0%	82.0%	1418	70	$0.50	$1.5
32	DeepSeek R1 DeepSeek	60.1	49.2%	86.0%	1355	35	$0.55	$2.2
33	DeepSeek V3.1 DeepSeek	59.3	46.0%	84.0%	1340	65	$0.14	$0.28
34	Qwen 3 Max Qwen	59.3	46.0%	86.0%	1340	55	$0.46	$1.8
35	o1 Mini OpenAI	58.5	45.0%	89.0%	1300	80	$3.0	$12.0
36	Grok 3 xAI	57.8	42.0%	82.0%	1402	65	$3.0	$15.0
37	Qwen 3 235B Qwen	56.8	42.0%	84.0%	1320	60	$0.20	$0.80
38	Codestral Mistral	55.9	40.0%	86.0%	1260	90	$0.30	$0.90
39	Claude 3.5 Sonnet Anthropic	55.3	49.0%	81.7%	1268	70	$3.0	$15.0
40	GPT-4o OpenAI	55.2	30.7%	87.2%	1280	143	$2.5	$10.0
41	GPT-4.1 Mini OpenAI	54.8	35.0%	80.0%	1250	160	$0.40	$1.6
42	o1 OpenAI	54	48.9%	89.0%	1360	35	$15.0	$60.0
43	Gemini 2.0 Flash Google	53.4	28.0%	76.0%	1240	220	$0.10	$0.40
44	Llama 4 Maverick Meta	53.2	35.0%	80.0%	1280	100	$0.25	$0.80
45	Claude Haiku 4.5 Anthropic	51.7	30.0%	78.0%	1220	180	$1.0	$5.0
46	Grok 3 Mini xAI	51.4	30.0%	78.0%	1260	130	$0.30	$0.50
47	Claude Fable 5 Anthropic	51.1	95.0%	—%	—	60	$10.0	$50.0
48	Qwen 3 32B Qwen	50.9	30.0%	79.0%	1250	120	$0.10	$0.30
49	GPT-5 Nano OpenAI	50.8	25.0%	72.0%	1200	250	$0.05	$0.40
50	Pixtral Large Mistral	50.2	35.0%	80.0%	1270	60	$2.0	$6.0
51	Gemini 2.5 Flash Lite Google	49.7	22.0%	70.0%	1230	240	$0.10	$0.40
52	GPT-4o Mini OpenAI	49.6	20.0%	78.0%	1220	200	$0.15	$0.60
53	Llama 4 Scout Meta	49.3	28.0%	75.0%	1240	140	$0.17	$0.50
54	Mistral Medium 3.1 Mistral	48.8	28.0%	77.0%	1250	110	$0.40	$2.0
55	Qwen 2.5 72B Qwen	47.9	28.0%	80.0%	1230	80	$0.12	$0.39
56	Command A Cohere	47.7	32.0%	78.0%	1250	70	$2.5	$10.0
57	Grok 2 xAI	46.9	28.0%	78.0%	1250	80	$2.0	$10.0
58	Claude 3.5 Haiku Anthropic	45.5	22.0%	74.0%	1180	170	$0.80	$4.0
59	Mistral Small 3.2 Mistral	43.7	20.0%	70.0%	1190	160	$0.10	$0.30
60	Llama 3.3 70B Meta	42.6	22.0%	72.0%	1210	90	$0.10	$0.30
61	GPT-4.1 Nano OpenAI	40.3	18.0%	65.0%	1120	200	$0.10	$0.40
62	Command R+ Cohere	38.9	20.0%	72.0%	1200	60	$2.5	$10.0
63	Nova Pro Amazon	38.7	18.0%	68.0%	1180	100	$0.80	$3.2
64	Claude 3 Opus Anthropic	34.2	22.0%	78.0%	1240	25	$15.0	$75.0
65	Command R Cohere	32.3	12.0%	62.0%	1130	90	$0.15	$0.60
66	GPT-4.5 OpenAI	31.6	38.0%	88.0%	1310	60	$75.0	$150.0
67	Nova Lite Amazon	31.1	10.0%	55.0%	1110	150	$0.06	$0.24
68	Reka Flash 3 Reka AI	30.4	—%	—%	—	136	$0.20	$0.80
69	Jamba 1.5 Mini AI21 Labs	25.7	—%	—%	—	35	$0.20	$0.40
70	Yi Lightning 01.AI	25	—%	—%	—	—	$0.14	$0.14
71	Sonar Perplexity	24.8	—%	—%	—	—	$1.0	$1.0
72	Sonar Pro Perplexity	24.7	—%	—%	—	50	$3.0	$15.0
73	Sonar Reasoning Pro Perplexity	24.2	—%	—%	—	22	$2.0	$8.0
74	Jamba 1.5 Large AI21 Labs	24	—%	—%	—	19	$2.0	$8.0
75	GLM-5 Zhipu AI	15	—%	—%	1456	55	—	—
76	GPT-OSS 120B OpenAI	15	—%	—%	—	339	—	—
77	MiniMax M2.5 MiniMax	14.6	—%	—%	1443	59	—	—
78	Doubao Seed 2.0 ByteDance	14	—%	—%	1474	—	—	—
79	GPT-OSS 20B OpenAI	13.7	—%	—%	—	312	—	—
80	Veo 3.1 Google	11.2	—%	—%	1401	—	—	—
81	Gemma 3 27B Google	10.6	—%	—%	1339	58	—	—
82	Sora 2 OpenAI	9.9	—%	—%	1368	—	—	—
83	Nova 2.0 Lite Amazon	9.5	—%	—%	—	221	—	—
84	Veo 3 Google	8.8	—%	—%	1340	—	—	—
85	Seedream 4.5 ByteDance	7.9	—%	—%	1316	—	—	—
86	Ministral 3 8B Mistral	7.2	—%	—%	—	172	—	—
87	MiMo V2 Flash Xiaomi	6.4	—%	—%	—	155	—	—
88	Phi-4 Mini Microsoft	6.1	—%	—%	—	150	—	—
89	Qwen 3 Next 80B Qwen	5.6	—%	—%	—	138	—	—
90	Phi-4 Microsoft	3.5	—%	—%	—	93	—	—
91	Gemma 3 12B Google	3.3	—%	—%	—	90	—	—
92	Ministral 3 14B Mistral	3	—%	—%	—	83	—	—
93	Ring Flash 2.0 InclusionAI	3	—%	—%	—	83	—	—
94	Nemotron 3 Nano NVIDIA	2.7	—%	—%	—	76	—	—
95	GLM-4.6V Zhipu AI	2.3	—%	—%	—	68	—	—
96	Step 2.5 Flash StepFun	2.3	—%	—%	—	67	—	—
97	K-EXAONE LG AI Research	2	—%	—%	—	62	—	—
98	Qwen 3 Coder 480B Qwen	1.9	—%	—%	—	60	—	—
99	Qwen 3 VL 235B Qwen	1.3	—%	—%	—	46	—	—
100	Kimi K2.5 Moonshot AI	1.2	—%	—%	—	45	—	—
101	Magistral Medium 1.2 Mistral	0.5	—%	—%	—	29	—	—
102	Phi-4 Reasoning Plus Microsoft	0	—%	—%	—	—	—	—
103	Gemma 3 4B Google	0	—%	—%	—	—	—	—
104	Mistral Large 3 Mistral	0	—%	—%	—	—	—	—
105	Kling 2.5 Turbo Kuaishou	0	—%	—%	—	—	—	—
106	DALL-E 3 OpenAI	0	—%	—%	—	—	—	—
107	Midjourney v7 Midjourney	0	—%	—%	—	—	—	—
108	Midjourney v6.1 Midjourney	0	—%	—%	—	—	—	—
109	Stable Diffusion 3.5 Stability AI	0	—%	—%	—	—	—	—
110	Flux 1.1 Pro Black Forest Labs	0	—%	—%	—	—	—	—
111	Flux 1.0 Dev Black Forest Labs	0	—%	—%	—	—	—	—
112	Imagen 4 Google	0	—%	—%	—	—	—	—
113	Ideogram 3.0 Ideogram	0	—%	—%	—	—	—	—
114	ERNIE 4.5 Baidu	0	—%	—%	—	—	—	—
115	ERNIE X1 Baidu	0	—%	—%	—	—	—	—
116	Hermes 4 70B Nous Research	0	—%	—%	—	—	—	—
117	Luma Ray 3 Luma AI	0	—%	—%	—	—	—	—
118	Runway Gen-4 Runway	0	—%	—%	—	—	—	—
119	Pika 2.5 Pika	0	—%	—%	—	—	—	—

Data sourced from Chatbot Arena, OpenRouter, and public benchmarks. Updated daily. Scores are dynamically computed based on your selected metrics.

How It Works

Choose a Persona or Build Your Own

Select a preset (Developer, Researcher, Business, etc.) that pre-selects relevant benchmarks and weights. Or customize everything from scratch.

See Real Data, Not Abstract Scores

Every metric shows the actual value — ELO 1410, $2.50/1M tokens, 65 tok/s. No normalization black box. The raw numbers are always visible.

Dynamic Ranking — Your Weights, Your Score

For each metric you select, we find the min/max across all models, normalize to 0-100, then compute a weighted composite. Missing data is excluded and weights renormalize automatically.

Share Your Rankings

Your exact configuration is encoded in the URL. Share it with your team or embed it — they'll see your exact ranking.

Data sourced from Chatbot Arena, OpenRouter, SWE-bench, and public research papers.

14 benchmarks across General, Coding, Math, Reasoning, Speed, Cost, and Context categories.

Frequently Asked Questions

What is the best AI model in 2026?

The best AI model depends on your use case. As of 2026, top contenders include Gemini 3.1 Pro, Claude Opus 4.6, and GPT-5.2 for general intelligence. For coding, Claude Opus 4.6 and GPT-5.2 lead on SWE-bench. For budget-conscious users, DeepSeek V3.2 and Gemini 2.5 Flash offer excellent performance per dollar. Use the AI Value Index to rank models based on YOUR priorities.

How do AI benchmarks work?

AI benchmarks are standardized tests that evaluate language models across specific capabilities. Common benchmarks include Chatbot Arena ELO (human preference voting), SWE-bench (real software engineering tasks), MMLU-Pro (knowledge and reasoning), GPQA Diamond (graduate-level science), and MATH (competition mathematics). Each benchmark tests a different aspect of model capability, and no single benchmark tells the whole story.

What is Chatbot Arena ELO?

Chatbot Arena ELO is a human preference ranking system where users compare AI model responses in blind head-to-head matchups. The ELO rating (borrowed from chess) reflects how often a model is preferred over others. Higher ELO means the model is more frequently preferred. It's considered one of the most reliable benchmarks because it uses real human judgment rather than automated scoring.

Which AI model is best for coding?

For coding tasks in 2026, the top models are Claude Opus 4.6 (80.8% SWE-bench), GPT-5.2 (80.0% SWE-bench), and Gemini 3.1 Pro (80.6% SWE-bench). For more affordable coding, GPT-5.1 Codex and Qwen 3 Coder offer strong performance at lower costs. The AI Value Index Developer persona pre-weights coding benchmarks to help you find the best fit.

Which AI model is cheapest?

The cheapest AI models by API pricing include GPT-5 Nano ($0.05/1M input), GPT-4.1 Nano ($0.10/1M input), Nova Lite ($0.06/1M input), and Gemini 2.5 Flash Lite ($0.10/1M input). For the best balance of quality and cost, DeepSeek V3.2 ($0.14/1M input) and Qwen 3 32B ($0.10/1M input) offer strong performance at budget prices.