Model rankings

Rank	Change	Model	Intelligence	Price $/M	Speed tok/s	First token
1	▲ 1	Inkling SmallOpenThinking Machines	32.2	$0.52	125	2.0 s
2	▼ 1	Inkling (xhigh)OpenThinking Machines	32.2	$1.8	72	2.6 s
3	—	Nemotron 3 Ultra 550B A55B (Reasoning)OpenNVIDIA	29.3	$1.1	166	2.3 s
4	—	Muse Glimmer (high)OpenMeta	24.4	$0.64	102	1.0 s
5	—	Gemma 4 31B (Reasoning)OpenGoogle · 1 more variant	22.2	$0.000	35	1.0 s
6	—	Gemma 4 26B A4B (Reasoning)OpenGoogle · 1 more variant	19.2	$0.18	—	—
7	—	Nemotron 3 Super 120B A12B (Reasoning)OpenNVIDIA	18.6	$0.35	141	2.0 s
8	▲ 1	Granite 4.2 30BOpenIBM	16.5	$0.28	59	1.2 s
9	▲ 1	Nemotron 3.5 LightningOpenNVIDIA	16.4	$0.095	292	0.6 s
10	▼ 2	gpt-oss-120b (high)OpenOpenAI · 1 more variant	15.6	$0.26	189	0.8 s
11	—	Gemma 4 12B (Reasoning)OpenGoogle · 1 more variant	15.6	$0.15	113	2.3 s
12	—	Apriel-v1.5-15B-ThinkerOpenServiceNow	15.0	$0.000	—	—
13	—	Apriel-v1.6-15B-ThinkerOpenServiceNow	14.3	$0.000	—	—
14	—	Granite 4.2 8BOpenIBM	13.9	$0.11	120	0.6 s
15	—	Nemotron Cascade 2 30B A3BOpenNVIDIA	11.5	—	—	—
16	▲ 4	Granite 4.2 3BOpenIBM	10.7	$0.053	205	0.5 s
17	▼ 1	gpt-oss-20b (high)OpenOpenAI · 1 more variant	9.1	$0.093	208	0.7 s
18	▼ 1	Nemotron 3 Nano Omni 30B A3B ReasoningOpenNVIDIA	9.0	$0.13	325	1.1 s
19	▼ 1	NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)OpenNVIDIA · 1 more variant	8.6	$0.088	166	1.4 s
20	▼ 1	Llama 4 MaverickOpenMeta	8.5	$0.42	86	0.9 s
21	—	DiffusionGemma 26B A4BOpenGoogle	7.6	—	—	—
22	—	Llama Nemotron Super 49B v1.5 (Reasoning)OpenNVIDIA · 1 more variant	6.6	$0.40	23	17 s
23	—	Llama 3.3 Nemotron Super 49B v1 (Reasoning)OpenNVIDIA · 1 more variant	6.4	—	—	—
24	—	Gemma 4 E4B (Reasoning)OpenGoogle · 1 more variant	6.4	$0.040	39	0.9 s
25	—	LFM2.5-2.6BOpenLiquid AI	5.3	$0.000	201	2.2 s
26	—	Llama 4 ScoutOpenMeta	4.6	$0.31	95	0.8 s
27	—	Gemma 4 E2B (Reasoning)OpenGoogle · 1 more variant	4.0	—	—	—
28	—	Llama 3.3 Instruct 70BOpenMeta	3.7	$0.67	91	1.6 s
29	—	Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)OpenNVIDIA	3.4	$0.90	52	2.4 s
30	—	NVIDIA Nemotron Nano 12B v2 VL (Reasoning)OpenNVIDIA · 1 more variant	3.3	$0.30	18	11 s
31	—	Granite 4.1 30BOpenIBM	3.2	—	—	—
32	—	NVIDIA Nemotron Nano 9B V2 (Reasoning)OpenNVIDIA · 1 more variant	3.2	$0.070	97	5.3 s
33	—	NVIDIA Nemotron 3 Nano 4BOpenNVIDIA	3.1	—	—	—
34	—	Llama 3.1 Nemotron Nano 4B v1.1 (Reasoning)OpenNVIDIA	2.9	—	—	—
35	—	Llama 3.1 Instruct 405BOpenMeta	2.9	—	—	—
36	—	LFM2.5-8B-A1BOpenLiquid AI	2.7	$0.000	—	—
37	—	Llama 3.1 Nemotron Instruct 70BOpenNVIDIA	2.1	$1.2	15	12 s
38	—	Llama 3.1 Instruct 8BOpenMeta	2.0	$0.028	142	0.8 s
39	—	Gemma 3 27B InstructOpenGoogle	2.0	$0.000	—	—
40	—	Llama 3.1 Instruct 70BOpenMeta	1.2	$0.56	65	1.6 s
41	—	Granite 4.1 8BOpenIBM	1.1	$0.063	101	0.8 s
42	▲ 14	Arctic InstructOpenSnowflake	1.0	—	—	—
42	▲ 18	DBRX InstructOpenDatabricks	1.0	—	—	—
42	▲ 3	Gemma 3 12B InstructOpenGoogle	1.0	$0.000	—	—
42	▲ 29	Gemma 3 1B InstructOpenGoogle	1.0	$0.000	—	—
42	▲ 23	Gemma 3 270MOpenGoogle	1.0	—	—	—
42	▲ 29	Gemma 3 4B InstructOpenGoogle	1.0	$0.000	—	—
42	▲ 29	Gemma 3n E2B InstructOpenGoogle	1.0	$0.000	—	—
42	▲ 29	Gemma 3n E4B InstructOpenGoogle	1.0	—	—	—
42	▲ 9	Gemma 3n E4B Instruct Preview (May '25)OpenGoogle	1.0	—	—	—
42	▲ 28	Granite 3.3 8B (Non-reasoning)OpenIBM	1.0	$0.085	16	26 s
42	▲ 26	Granite 4.0 1BOpenIBM	1.0	—	—	—
42	▲ 29	Granite 4.0 350MOpenIBM	1.0	—	—	—
42	▲ 22	Granite 4.0 H 1BOpenIBM	1.0	—	—	—
42	▲ 29	Granite 4.0 H 350MOpenIBM	1.0	—	—	—
42	▲ 4	Granite 4.0 H SmallOpenIBM	1.0	$0.11	26	14 s
42	▲ 24	Granite 4.0 MicroOpenIBM	1.0	—	—	—
42	▲ 7	Granite 4.1 3BOpenIBM	1.0	—	—	—
42	▲ 29	LFM2 1.2BOpenLiquid AI	1.0	—	—	—
42	▲ 20	LFM2 2.6BOpenLiquid AI	1.0	—	—	—
42	▲ 5	LFM2 24B A2BOpenLiquid AI	1.0	—	—	—
42	▲ 27	LFM2 8B A1BOpenLiquid AI	1.0	—	—	—
42	▲ 21	LFM2.5-1.2B-InstructOpenLiquid AI	1.0	—	—	—
42	▲ 19	LFM2.5-1.2B-ThinkingOpenLiquid AI	1.0	—	—	—
42	▲ 29	LFM2.5-VL-1.6BOpenLiquid AI	1.0	$0.000	—	—
42	▲ 17	Llama 2 Chat 13BOpenMeta	1.0	—	—	—
42	▲ 16	Llama 2 Chat 70BOpenMeta	1.0	—	—	—
42	▲ 11	Llama 2 Chat 7BOpenMeta	1.0	$0.10	—	—
42	▲ 13	Llama 3 Instruct 70BOpenMeta	1.0	$1.2	—	—
42	▲ 29	Llama 3 Instruct 8BOpenMeta	1.0	$0.070	—	—
42	▲ 15	Llama 3.2 Instruct 11B (Vision)OpenMeta	1.0	$0.35	21	1.2 s
42	▲ 29	Llama 3.2 Instruct 1BOpenMeta	1.0	—	—	—
42	▲ 12	Llama 3.2 Instruct 3BOpenMeta	1.0	—	—	—
42	▲ 1	Llama 3.2 Instruct 90B (Vision)OpenMeta	1.0	—	—	—
42	▲ 25	Llama 65BOpenMeta	1.0	—	—	—
42	▲ 8	Phi-3 Mini Instruct 3.8BOpenMicrosoft	1.0	—	—	—
42	▲ 6	Phi-4OpenMicrosoft	1.0	$0.22	40	2.5 s
42	▲ 2	Phi-4 Mini InstructOpenMicrosoft	1.0	$0.000	45	0.8 s
42	▲ 10	Phi-4 Multimodal InstructOpenMicrosoft	1.0	$0.000	17	0.8 s
42	—	R1 1776OpenPerplexity	1.0	—	—	—

30 more models

Intelligence over time

Each maker's best model at the time on the Artificial Analysis Intelligence Index, by the month it was released.

Gained most · Anthropic +57.2Highest now · Anthropic 64.6Gap between highest and lowest 18.1 ptsTime range Nov 2025 – Nov 2026

64.6Anthropic60.0Google56.9OpenAI50.1Meta46.5Kimi