Intelligence per dollar

Name: LLM intelligence per dollar
Creator: AI Calculator Pro
License: https://creativecommons.org/licenses/by/4.0/

The smartest model is rarely the one you should ship. This board ranks every major LLM by community Arena score and by intelligence per dollar, and flags the models on the price/quality frontier — the ones nothing else beats on both quality and cost. Pick an arena, then sort by what matters for your task.

Find the cheapest model for my task Can I downgrade & save?

Decision-ready picks

Skip the leaderboard — here’s the one model that best answers each common question (overall arena).

Best overall

Kimi K3

Moonshot AI · Arena 1473

The highest-scoring model we track for this arena — the quality ceiling.

Best value

Amazon Nova Micro

AWS Bedrock · 20702 Elo per $/1M

The most Arena score per dollar — punches far above its price.

Cheapest that clears the bar

Amazon Nova Micro

AWS Bedrock · $0.061/1M blended

The lowest price among models still within 15% of the top Arena score.

Largest useful context

Llama 4 Scout

Meta · 10M tokens

The biggest context window among genuinely capable models — for long docs and codebases.

Best for coding

Kimi K3

Moonshot AI · Coding Arena 1679

Top of the coding Arena — the safest pick for agents and multi-file edits.

Price vs performance

Every scored model plotted by price and Arena score. The best value sits toward the top-left.

Price vs quality — lower and higher is better. Highlighted points are on the price/quality frontier (nothing we track beats them on both). Price uses a 3:1 blended $/1M on a log scale.

11 of 127 models are on the price/quality frontier (marked Best value). An Est. tag means the score is inherited from the base model (a regional/creator-prefixed hosting duplicate). Click a column to sort; click again to flip.


1	Kimi K3Best value	1473	$6.00	246	Moonshot AI
2	GPT-5.5	1465	$11.25	130	OpenAI
3	GPT-5.5Est.	1465	$12.38	118	AWS Bedrock
4	Gemini 2.5 Pro	1463	$3.44	426	Google
5	GLM-5.2Best value	1463	$2.15	680	Z.ai (Zhipu AI)
6	GPT-5.4	1461	$5.63	260	OpenAI
7	GPT-5.4Est.	1461	$6.19	236	AWS Bedrock
8	GPT-5.6 Sol	1455	$11.25	129	OpenAI
9	GPT-5.6 SolEst.	1455	$11.25	129	AWS Bedrock
10	Claude Fable 5	1454	$20.00	73	Anthropic
11	Claude Fable 5Est.	1454	$20.00	73	AWS Bedrock
12	Claude Opus 4.8	1453	$10.00	145	Anthropic
13	Claude Opus 4.8Est.	1453	$10.00	145	AWS Bedrock
14	Grok 4.5	1451	$3.00	484	xAI
15	Claude Opus 4.6	1449	$10.00	145	Anthropic
16	Claude Opus 4.5 (latest)	1449	$10.00	145	Anthropic
17	Claude Opus 5 (JP)Est.	1449	$10.00	145	AWS Bedrock
18	Claude Opus 5	1449	$10.00	145	Anthropic
19	GLM-4.6Best value	1448	$1.00	1448	Z.ai (Zhipu AI)
20	Claude Sonnet 4.6	1445	$6.00	241	Anthropic
21	MiMo-V2.5-ProBest value	1445	$0.544	2657	Xiaomi
22	AU Anthropic Claude Sonnet 4.6Est.	1445	$6.60	219	AWS Bedrock
23	Hunyuan Hy3Best value	1444	$0.350	4126	Tencent
24	Qwen3.7 Max	1444	$3.75	385	Alibaba
25	Claude Sonnet 5	1440	$4.00	360	Anthropic
26	Claude Opus 4.7 (US)Est.	1440	$10.00	144	AWS Bedrock
27	Claude Sonnet 5 (Global)Est.	1440	$4.00	360	AWS Bedrock
28	Claude Opus 4.7	1440	$10.00	144	Anthropic
29	Qwen3.5 397B-A17B	1439	$1.35	1066	Alibaba
30	Gemini 3.5 Flash Lite	1439	$0.850	1693	Google
31	Gemini 3.6 Flash	1439	$3.00	480	Google
32	GLM-5	1438	$1.55	928	Z.ai (Zhipu AI)
33	Claude Sonnet 4.5 (latest)	1438	$6.00	240	Anthropic
34	GLM-5Est.	1438	$1.55	928	AWS Bedrock
35	Kimi K2.6	1435	$1.71	838	Moonshot AI
36	MiMo-V2-Pro	1435	$0.544	2639	Xiaomi
37	DeepSeek V4 Pro	1434	$0.544	2637	DeepSeek
38	Gemini 3.5 Flash	1434	$3.38	425	Google
39	Qwen3.7 Plus	1433	$1.13	1274	Alibaba
40	GLM-4.7	1432	$1.00	1432	Z.ai (Zhipu AI)
41	GLM-4.7Est.	1432	$1.00	1432	AWS Bedrock
42	DeepSeek V4 FlashBest value	1430	$0.175	8171	DeepSeek
43	GLM-4.5	1429	$1.00	1429	Z.ai (Zhipu AI)
44	GLM-5.1	1429	$2.15	665	Z.ai (Zhipu AI)
45	MiMo-V2.5	1427	$0.175	8154	Xiaomi
46	GLM-5V-Turbo	1424	$1.90	749	Z.ai (Zhipu AI)
47	Mistral Large 3	1423	$0.750	1897	Mistral AI
48	o3	1422	$3.50	406	OpenAI
49	GPT-5.1	1422	$3.44	414	OpenAI
50	MiMo-V2-Omni	1422	$0.175	8126	Xiaomi
51	MiniMax-M3	1421	$0.525	2707	MiniMax
52	Kimi K2 Thinking Turbo	1421	$2.86	496	Moonshot AI
53	DeepSeek V3.2	1420	$0.315	4508	DeepSeek
54	Kimi K2.5	1420	$1.20	1183	Moonshot AI
55	Kimi K2.5Est.	1420	$1.20	1183	AWS Bedrock
56	Qwen3.6 Plus	1418	$1.13	1260	Alibaba
57	Qwen3.5 122B-A10B	1418	$1.10	1289	Alibaba
58	Gemini 2.5 Flash	1417	$0.850	1667	Google
59	Claude Opus 4.1 (latest)	1417	$30.00	47	Anthropic
60	Gemini 3.1 Flash Lite	1415	$0.563	2516	Google
61	GPT-5.4 mini	1413	$1.69	837	OpenAI
62	GPT-5.2	1412	$4.81	293	OpenAI
63	Muse Spark 1.1	1410	$2.00	705	Meta
64	Qwen3.5 27B	1408	$0.825	1707	Alibaba
65	GPT-5	1405	$3.44	409	OpenAI
66	MiniMax-M2.7	1405	$0.525	2676	MiniMax
67	Qwen3 Max	1404	$2.40	585	Alibaba
68	Step 3.5 FlashBest value	1404	$0.150	9360	StepFun
69	Qwen/Qwen3-VL-235B-A22B-InstructEst.	1401	$0.600	2335	AWS Bedrock
70	Qwen3-VL 235B-A22B	1401	$1.22	1144	Alibaba
71	Amazon Nova Premier	1400	$5.00	280	AWS Bedrock
72	Grok 4.3	1400	$1.56	896	xAI
73	Qwen3.5 35B-A3B	1396	$0.688	2031	Alibaba
74	MiMo-V2-Flash	1395	$0.175	7971	Xiaomi
75	Claude Haiku 4.5	1394	$2.00	697	Anthropic
76	Qwen3-Next 80B-A3B Instruct	1392	$0.875	1591	Alibaba
77	MiniMax-M2.1	1391	$0.525	2650	MiniMax
78	MiniMax M2.1Est.	1391	$0.525	2650	AWS Bedrock
79	GLM-4.5-Air	1383	$0.425	3254	Z.ai (Zhipu AI)
80	o4-mini	1382	$1.93	718	OpenAI
81	GLM-4.6V	1375	$0.450	3056	Z.ai (Zhipu AI)
82	GPT-5 Mini	1374	$0.688	1999	OpenAI
83	GPT-5.4 nano	1373	$0.463	2969	OpenAI
84	GPT-4.1	1372	$3.50	392	OpenAI
85	Kimi K2 Thinking	1371	$1.08	1275	Moonshot AI
86	Kimi K2 ThinkingEst.	1371	$1.08	1275	AWS Bedrock
87	Qwen3-Next 80B-A3B (Thinking)	1368	$1.88	730	Alibaba
88	Qwen3 235B-A22B	1366	$1.22	1115	Alibaba
89	Command A	1365	$4.38	312	Cohere
90	Llama 4 Maverick	1360	$0.378	3603	Meta
91	MiniMax-M2.5	1359	$0.525	2589	MiniMax
92	MiniMax M2.5Est.	1359	$0.525	2589	AWS Bedrock
93	Qwen3-Coder 480B-A35B Instruct	1356	$3.00	452	Alibaba
94	GPT-4o	1355	$4.38	310	OpenAI
95	Gemini 2.0 Flash	1354	$0.175	7737	Google
96	GLM-4.7-FlashBest value	1353	$0.145	9331	Z.ai (Zhipu AI)
97	o1	1353	$26.25	52	OpenAI
98	GLM-4.7-FlashEst.	1353	$0.153	8872	AWS Bedrock
99	Mistral Medium 3.1	1350	$0.800	1688	Mistral AI
100	Amazon Nova Pro	1348	$1.40	963	AWS Bedrock
101	GPT-4.1 Mini	1345	$0.700	1921	OpenAI
102	MiniMax-M2	1342	$0.525	2556	MiniMax
103	MiniMax M2Est.	1342	$0.525	2556	AWS Bedrock
104	Devstral 2	1340	$0.525	2552	Mistral AI
105	Qwen3 32B	1340	$1.22	1094	Alibaba
106	GLM-4.5V	1334	$0.900	1482	Z.ai (Zhipu AI)
107	Llama 4 Scout	1332	$0.168	7952	Meta
108	Gemini 2.5 Flash-Lite	1330	$0.175	7600	Google
109	Qwen Plus	1327	$0.600	2212	Alibaba
110	Mistral Small 4	1325	$0.262	5048	Mistral AI
111	Step 2 (16K)	1321	$8.02	165	StepFun
112	GPT-5 NanoBest value	1320	$0.138	9600	OpenAI
113	o3-mini	1319	$1.93	685	OpenAI
114	Llama 3.3 70B	1315	$0.600	2192	Meta
115	GPT-4o mini	1310	$0.262	4990	OpenAI
116	GPT-4.1 Nano	1300	$0.175	7429	OpenAI
117	Amazon Nova LiteBest value	1298	$0.105	12362	AWS Bedrock
118	Qwen Max	1282	$2.80	458	Alibaba
119	Llama-3.3-70B-Instruct	1275	$0.198	6456	Meta
120	Qwen2.5 72B Instruct	1269	$2.45	518	Alibaba
121	Amazon Nova MicroBest value	1268	$0.061	20702	AWS Bedrock
122	Command R7B	1262	$0.066	19230	Cohere
123	Phi 4	1217	$0.088	13909	Microsoft
124	Command R+	1204	$4.38	275	Cohere
125	GPT-4	1186	$37.50	32	OpenAI
126	Command R	1163	$0.262	4430	Cohere
127	GPT-3.5-turbo	1094	$0.750	1459	OpenAI

Intelligence scores are community Arena ratings from LMArena / arena.ai, used under CC BY 4.0. Snapshot last refreshed 28 July 2026. Scores are a relative signal, not an absolute measure of capability. A Measured score was voted on directly; an Estimated score is inherited from an identical base model (a regional/creator-prefixed hosting duplicate). See our methodology.

Download / cite this dataset

The full ranking is available as open data — reuse it with attribution and a link back.

Download JSON Download CSV

Cite as: AI Calculator Pro, “LLM intelligence per dollar”, https://aicalculatorpro.com/intelligence/ (snapshot 2026-07-28). Underlying scores LMArena / arena.ai, CC BY 4.0.

FAQ

What does the Arena score mean?

It is a community Elo rating from blind, head-to-head votes: users compare two anonymous model answers and pick the better one. A higher score means the model is preferred more often. It measures human preference, not correctness on any single benchmark.

What is 'intelligence per dollar'?

We divide a model's Arena score by its blended price (a 3:1 input:output blend, $ per 1M tokens). It surfaces models that punch above their price. Use it as a starting point, then confirm the model clears the quality bar your task actually needs.

Why isn't every model listed?

We only show a score when we can confidently map an Arena entry to a model we track. Models without a confident match carry no score rather than a wrong one. Scores come from LMArena / arena.ai and were last refreshed 2026-07-28.

What do the “Measured” and “Estimated” labels mean?

Measured means the Arena leaderboard voted on that exact model. Estimated means the model is a regional or creator-prefixed hosting listing of a base model (for example an AWS Bedrock “jp.anthropic.…” profile) — the weights are identical, so we inherit the base model's score, but Arena never voted on that specific listing. We label it rather than hide it, so you always know whether a score is direct or inherited.

Cite / link to this page

You’re welcome to reference this page and its figures with attribution and a link back.

https://aicalculatorpro.com/intelligence/

<a href="https://aicalculatorpro.com/intelligence/">Intelligence per dollar — LLM rankings — AI Calculator Pro</a>