Fast and good

At least 100 tokens per second with an index above 55.

25 models qualify · data as of 9 Sept 2026

1. 1Muse Spark 1.164.9
2. 2Gemini 3.7 Flash64.4
3. 3Gemini 3.5 Flash63.9
4. 4Muse Spark 1.363.9
5. 5Gemini 3.1 Pro63.9
6. 6Gemini 3.8 Flash63.5
7. 7GPT-5.463.4
8. 8Muse Spark 1.262.6
9. 9GPT-5.3 Codex62.4
10. 10Gemini 3 Flash61.5
11. 11Gemini 3.6 Flash60.7
12. 12Grok 4.360.7
13. 13GPT-5.5 Instant60.4
14. 14o360.3
15. 15GPT-5.159.4

SearchAll providersAlibabaDeepSeekGoogleMetaMiniMaxNVIDIAOpenAIxAIOpen weights25 of 25

#	Model	Index ↓	Answer ↕	Price $/M ↕	Speed tok/s ↕
1	Muse Spark 1.1Meta	64.9	194 tok/s	$2.00	194
2	Gemini 3.7 FlashmediumGoogle	64.4	282 tok/s	$1.50	282
3	Gemini 3.5 FlashmediumGoogle	63.9	209 tok/s	$3.38	209
4	Muse Spark 1.3xhighMeta	63.9	193 tok/s	$2.00	193
5	Gemini 3.1 ProGoogle	63.9	110 tok/s	$4.50	110
6	Gemini 3.8 FlashmediumGoogle	63.5	273 tok/s	$1.50	273
7	GPT-5.4OpenAI	63.4	130 tok/s	$5.63	130
8	Muse Spark 1.2Meta	62.6	219 tok/s	$2.00	219
9	GPT-5.3 CodexOpenAI	62.4	126 tok/s	$4.81	126
10	Gemini 3 FlashthinkingGoogle	61.5	200 tok/s	$1.13	200
11	Gemini 3.6 FlashGoogle	60.7	190 tok/s	$1.50	190
12	Grok 4.3mediumxAI	60.7	113 tok/s	$1.56	113
13	GPT-5.5 InstantOpenAI	60.4	135 tok/s	$11.25	135
14	o3OpenAI	60.3	139 tok/s	$3.50	139
15	GPT-5.1OpenAI	59.4	105 tok/s	$3.44	105
16	DeepSeek V4 FlashhighDeepSeekopen ↗	59.3	118 tok/s	$0.168	118
17	GPT-5.6 LunaxhighOpenAI	59.3	106 tok/s	$0.450	106
18	Qwen3 7maxAlibaba	58.9	199 tok/s	$3.75	199
19	Nemotron 3 Ultra 550B A55BNVIDIAopen ↗	58.3	156 tok/s	$1.00	156
20	Grok 4.20xAI	57.9	104 tok/s	$1.56	104
21	Deepseek v4 Flash VisionDeepSeek	57.4	119 tok/s	$0.175	119
22	MiniMax-M3MiniMaxopen ↗	56.9	100 tok/s	$0.525	100
23	Qwen3.5 122B-A10BAlibabaopen ↗	56	129 tok/s	$1.10	129
24	GPT-5 miniOpenAI	55.8	103 tok/s	$0.688	103
25	Muse Glimmer (high)Metaopen	55.1	104 tok/s	$0.638	104