Value: quality per dollar

BenchLeader Index divided by the blended price per million tokens (3 input : 1 output). Higher means more quality for the money.

As of 4 Oct 2026, BenchLeader ranks Gemma 4 E4B first on the Value: quality per dollar board at 1115.0, ahead of Granite 4.2 3B (849.1) and Gemma 4 E2B (834.0).

278 models · data as of 4 Oct 2026

Model sizeAll 83Tiny 2Small 11Medium 6Large 28

• OpenAI13
• Alibaba11
• Anthropic7
• Google7
• Mistral AI7
• DeepSeek6
• NVIDIA4
• Zhipu AI4
• Moonshot AI3
• SpaceXAI3
• Xiaomi3
• IBM2
• Meta2
• Ant Group1
• AApodex1
• Arcee AI1
• CCeleris1
• Inception1

The dashed staircase is the efficiency frontier: at each price, the best index anyone reaches. A model above and left of the rest gives you more quality per dollar of real work. Size bands and run costs come from Artificial Analysis, which is the only source that publishes what a fixed batch of work actually cost.

1. 1Gemma 4 E4B (thinking)1115.0
2. 2Granite 4.2 3B849.1
3. 3Gemma 4 E2B (thinking)834.0
4. 4Qwen3.5-4B (thinking)778.3
5. 5Sarvam 30B (high)776.0
6. 6LFM2 24B A2B771.7
7. 7Llama 3.1 8B684.9
8. 8Mercury 2.5683.8
9. 9Mistral Small 3681.0
10. 10Nova Micro647.5
11. 11deepseek-r1-0528-qwen3-8b555.9
12. 12Gemma 3 4B552.4
13. 13Sarvam 105B (high)552.1
14. 14Nemotron 3 Nano 30B A3B (thinking)530.7
15. 15Nemotron 3.5 Lightning502.1

SearchAll providersAi2AI21 LabsAlibabaAmazonAnt GroupAnthropicApodexArcee AIBaiduByteDanceCelerisCohereDeep CogitoDeepSeekGoogleIBMInceptionKwaiKATLiquid AIMeituanMetaMicrosoftMiniMaxMistral AIMoonshot AIMultiverse ComputingNex AGINous ResearchNVIDIAOpenAIRekaSarvamSpaceXAIStepFunSwiss AIThinking MachinesUpstageXiaomiZhipu AIOpen weightsNew models278 of 278

#	Model	Index per $ ↓	Index ↕	Blended $/M ↕
1	Gemma 4 E4BthinkingGoogleopen ↗	1115.0	44.6	$0.040
2	Granite 4.2 3BIBMopen ↗	849.1	45.0	$0.053
3	Gemma 4 E2BthinkingGoogleopen ↗	834.0	41.7	$0.050
4	Qwen3.5-4BthinkingAlibabaopen ↗	778.3	46.7	$0.060
5	Sarvam 30BhighSarvamopen	776.0	38.8	$0.050
6	LFM2 24B A2BLiquid AIopen ↗	771.7	40.9	$0.053
7	Llama 3.1 8BMetaopen ↗	684.9	36.3	$0.053
8	Mercury 2.5Inception	683.8	46.5	$0.068
9	Mistral Small 3Mistral AIopen ↗	681.0	39.5	$0.058
10	Nova MicroAmazon	647.5	39.5	$0.061
11	deepseek-r1-0528-qwen3-8bDeepSeekopen ↗	555.9	37.8	$0.068
12	Gemma 3 4BGoogleopen ↗	552.4	34.8	$0.063
13	Sarvam 105BhighSarvamopen	552.1	40.3	$0.073
14	Nemotron 3 Nano 30B A3BthinkingNVIDIAopen ↗	530.7	46.7	$0.088
15	Nemotron 3.5 LightningNVIDIAopen ↗	502.1	47.7	$0.095
16	Gemma 3 12BGoogleopen ↗	497.3	37.3	$0.075
17	Ling 3.0 Flash VLAnt Groupopen ↗	485.2	55.8	$0.115
18	Llama 3-8BMetaopen	483.6	35.3	$0.073
19	Ling 3.0 Flash FinAnt Groupopen	473.9	54.5	$0.115
20	Granite 3.3 8Bno reasoningIBMopen ↗	462.4	39.3	$0.085
21	Ling 3.0 FlashAnt Groupopen ↗	460.9	53.0	$0.115
22	Llama 3.2 1B InstructMetaopen ↗	459.2	32.6	$0.071
23	Qwen3.5-9BthinkingAlibabaopen ↗	434.5	49.1	$0.113
24	Granite 4.2 8BIBMopen ↗	429.6	46.4	$0.108
25	GPT-5 nanoOpenAI	405.1	55.9	$0.138
26	nvidia-nemotron-nano-9b-v2thinkingNVIDIAopen	403.9	41.6	$0.103
27	Ministral 8BMistral AIopen	381.0	38.1	$0.100
28	Granite 4.0 H SmallIBMopen ↗	373.1	40.3	$0.108
29	Nova LiteAmazon	372.4	39.1	$0.105
30	Ministral 3 3BMistral AIopen	371.0	37.1	$0.100
31	Step 3.5 FlashStepFunopen ↗	360.0	54.0	$0.150
32	MiMo-V2-FlashthinkingXiaomiopen ↗	351.3	52.7	$0.150
33	gpt-oss-20bhighOpenAIopen ↗	346.1	44.3	$0.128
34	Llama 2-7BMetaopen ↗	340.0	34.0	$0.100
35	MiMo-V2.6-FlashXiaomiopen ↗	337.7	59.1	$0.175
36	GLM 4.7 FlashZhipu AIopen ↗	337.3	51.6	$0.153
37	Gemma 4 12BthinkingGoogleopen ↗	337.3	50.6	$0.150
38	Olmo 3 7B ThinkthinkingAi2open	319.2	39.9	$0.125
39	MiMo-V2.5Xiaomiopen ↗	313.7	54.9	$0.175
40	DeepSeek V4 Flash 0731highDeepSeekopen ↗	310.9	54.4	$0.175
41	Qwen3 30B A3B 2507Alibabaopen ↗	305.3	45.8	$0.150
42	Apertus 8B InstructSwiss AIopen	300.8	37.6	$0.125
43	Llama 3.2 3B InstructMetaopen ↗	300.0	36.0	$0.120
44	Qwen3.5 FlashAlibaba	298.9	52.3	$0.175
45	GPT-6 LunamaxOpenAI	296.5	59.3	$0.200
46	Gemma 4 26B A4BGoogleopen ↗	285.9	56.6	$0.198
47	Devstral Small 2505Mistral AIopen	283.3	42.5	$0.150
48	Mistral Small 3.2Mistral AIopen ↗	280.0	42.0	$0.150
49	Gemini 2.5 Flash-LiteGoogle	276.0	48.3	$0.175
50	Gemini 2.5 Flash Lite 09thinkingGoogle	274.3	48.0	$0.175
51	Phi-4-MiniMicrosoftopen	274.0	35.9	$0.131
52	GLM 5.3 FlashZhipu AIopen ↗	262.2	62.4	$0.238
53	Gemma 3 27BGoogleopen ↗	260.7	39.1	$0.150
54	Qwen3.8-Flash-NextAlibabaopen ↗	260.4	59.9	$0.230
55	Ministral 3 8BMistral AIopen	257.3	38.6	$0.150
56	Gemma 4 31BGoogleopen ↗	255.1	52.3	$0.205
57	DeepSeek V4.1 FlashmaxDeepSeekopen ↗	235.9	61.8	$0.262
58	DeepSeek V4 FlashmaxDeepSeekopen ↗	231.7	60.7	$0.262
59	Mistral Small 3.1Mistral AIopen	230.4	36.4	$0.158
60	Qwen3 30B A3BAlibabaopen ↗	226.0	48.6	$0.215

Show 100 more

Cite as: BenchLeader, “Value: quality per dollar”, https://www.benchleader.com/leaderboards/value, data as of 4 Oct 2026.