Response time: seconds to a complete answer

Median seconds from request to the last token, reasoning included (Artificial Analysis). Time to first answer tells you when output starts; this tells you when it ends.

As of 9 Nov 2026, BenchLeader ranks Celeris-1 first on the Response time: seconds to a complete answer board at 3.9 s, ahead of Mercury 2.5 (5.6 s) and GPT-4.1 nano (6.3 s).

415 models · data as of 9 Nov 2026

Model sizeAll 108Tiny 0Small 7Medium 4Large 26Join effort settings

Copy linkDownload imageShow as table

The purple line 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, task costs and token counts come from Artificial Analysis, which is the only source that publishes what a fixed batch of work actually cost and how much a model wrote for it.

1. 1Celeris-13.9 s
2. 2Mercury 2.55.6 s
3. 3GPT-4.1 nano6.3 s
4. 4Gemini 3 Flash (none)6.8 s
5. 5Nova Micro7.3 s
6. 6Gemini 2.5 Flash-Lite (none)7.5 s
7. 7Gemini 3.1 Flash Lite7.9 s
8. 8GPT-5.4 Mini (none)8.6 s
9. 9NVIDIA Nemotron Nano 9B v2 (thinking)9.6 s
10. 10Nova Lite9.7 s
11. 11GPT-4.1 mini (high)10.2 s
12. 12Nemotron 3 Nano 30B A3B (none)10.6 s
13. 13Nemotron Nano 12B v2 VL (none)10.6 s
14. 14Mercury 210.8 s
15. 15GPT-6 Sol (none)10.9 s

SearchAll providersAlibabaAmazonAnt GroupAnthropicApodexByteDanceCelerisCohereDeepSeekGoogleIBMInceptionMBZUAIMeituanMetaMicrosoftMiniMaxMistral AIMoonshot AINous ResearchNVIDIAOpenAIReflectionSapiens AISarvamSpaceXAIStepFunTencentThinking MachinesUpstageXiaomiZhipu AIOpen weightsNew modelsBest per model (205)205 of 205

#	Model	seconds ↓	Thinking ↕	Index ↕
1	Kimi K3.1 ProMoonshot AIopen ↗	6.7 min	2.7 min	61.2
2	Qwen3.8 MaxmaxAlibaba	4.7 min	2.0 min	58.9
3	Qwen3.8 2.4T A95BAlibabaopen ↗	4.6 min	1.9 min	51.6
4	Command ACohereopen ↗	3.5 min	–	42.1
5	Qwen3.8-Flash-NextAlibabaopen ↗	3.3 min	1.5 min	54.4
6	Claude Fable 5maxAnthropic	3.2 min	–	65.3
7	Ling Flash 2.0Ant Groupopen	3.2 min	–	46.8
8	GLM 5.3 FlashZhipu AIopen ↗	3.1 min	1.4 min	56.8
9	MiniMax-M1MiniMaxopen	3.1 min	–	49.6
10	MiniMax M3MiniMaxopen ↗	2.7 min	1.1 min	56.3
11	Phi-4-MiniMicrosoftopen ↗	2.4 min	–	–
12	Seed-OSS-36B-InstructByteDanceopen ↗	2.2 min	45.3 s	49.2
13	GPT-5.2-CodexxhighOpenAI	2.2 min	–	–
14	Llama 3.1 70BMetaopen ↗	2.2 min	–	40.8
15	DeepSeek R1 0528DeepSeekopen ↗	2.2 min	1.2 min	49.3
16	MiMo-V2.6-FlashXiaomiopen ↗	2.1 min	37.6 s	54.9
17	MiniMax M2.7MiniMaxopen ↗	2.0 min	26.6 s	56.5
18	MiniMax-M2MiniMaxopen ↗	2.0 min	28.3 s	52.6
19	Qwen3.7 PlusAlibaba	1.9 min	41.0 s	57.7
20	Kimi K2.6no reasoningMoonshot AIopen ↗	1.9 min	–	–
21	Qwen3.5-4Bno reasoningAlibabaopen ↗	1.9 min	–	44.0
22	Solar Mini 4Upstage	1.8 min	57.1 s	–
23	o4-minihighOpenAI	1.8 min	–	50.3
24	Qwen3.6 PlusAlibaba	1.8 min	42.5 s	59.1
25	Gemini 4 ArgonhighGoogle	1.8 min	–	65.7
26	MiniMax M3.1 FlashMiniMaxopen ↗	1.8 min	44.1 s	53.9
27	Deepseek v4 Flash VisionmaxDeepSeek	1.7 min	53.7 s	–
28	Hy3Tencentopen ↗	1.7 min	43.2 s	55.8
29	MiMo-V2.6-ProXiaomiopen ↗	1.7 min	22.0 s	57.1
30	Granite 4.2 8BIBMopen ↗	1.6 min	33.7 s	41.1
31	GPT-6 AstralowOpenAI	1.6 min	–	62.2
32	GLM-4.6thinkingZhipu AIopen ↗	1.5 min	28.8 s	51.4
33	Solar Pro 4Upstage	1.5 min	40.4 s	49.5
34	Ling 3.1 FlashAnt Groupopen	1.5 min	33.3 s	50.8
35	LongCat 2.0Meituanopen ↗	1.5 min	29.5 s	–
36	Claude Opus 4.8maxAnthropic	1.5 min	–	60.2
37	Claude Opus 4.6no reasoningAnthropic	1.5 min	–	58.0
38	MiniMax M2.5MiniMaxopen ↗	1.5 min	15.4 s	53.6
39	MiniMax-M2.1MiniMaxopen ↗	1.5 min	15.0 s	55.0
40	Kimi K2.7 CodeMoonshot AIopen ↗	1.5 min	28.3 s	54.2
41	NewInkling 2 SmallThinking Machines	1.4 min	27.4 s	58.3
42	GPT-5.3-CodexxhighOpenAI	1.4 min	–	62.9
43	Claude Opus 5lowAnthropic	1.4 min	–	57.0
44	DeepSeek V4 FlashmaxDeepSeekopen ↗	1.4 min	40.1 s	57.0
45	Llama 3.2 Instruct 11B (Vision)Metaopen ↗	1.3 min	–	35.6
46	Inkling SmallThinking Machinesopen ↗	1.3 min	26.5 s	52.1
47	Mistral Medium 3.5Mistral AIopen	1.3 min	20.0 s	46.6
48	Kimi K2Moonshot AIopen ↗	1.2 min	–	51.0
49	Gemma 3 27BGoogleopen ↗	1.2 min	–	39.9
50	Step 3.5 Flash 0202StepFunopen	1.2 min	27.8 s	–
51	Solar Pro 3Upstage	1.1 min	21.6 s	40.9
52	Ministral 3 14BMistral AIopen ↗	1.1 min	–	39.9
53	Mistral Medium 3Mistral AI	1.1 min	–	44.3
54	Kimi K2.5no reasoningMoonshot AIopen ↗	1.1 min	–	–
55	Grok 3 minithinkingSpaceXAI	1.0 min	22.6 s	–
56	DeepSeek V4 Flash 0420maxDeepSeekopen	59.6 s	15.2 s	–
57	Claude Fable 5.1lowAnthropic	58.9 s	–	58.7
58	GPT-6.1 SollowOpenAI	57.8 s	–	60.2
59	Step 3.5 FlashStepFunopen ↗	57.1 s	23.5 s	56.4
60	Kimi K3lowMoonshot AIopen ↗	56.0 s	3.1 s	53.7

Show 100 more

Cite as: BenchLeader, “Response time: seconds to a complete answer”, data as of 9 Nov 2026, https://www.benchleader.com/leaderboards/response-time.