Time per task: summarise a 30-page report

12,000 tokens of document in, a 600-token summary out. Time = time to first answer + output tokens ÷ output speed.

426 models · data as of 9 Sept 2026

1. 1Gemini 2.5 Flash-Lite2.4 s
2. 2Gemini 3.7 Flash2.6 s
3. 3Nova Micro3.1 s
4. 4Gemini 2.5 Flash3.5 s
5. 5Gemini 3 Flash3.8 s
6. 6Gemini 3 Flash3.8 s
7. 7Gemini 3 Flash3.8 s
8. 8Ministral 3 3B3.9 s
9. 9Gemini 3.5 Flash4.2 s
10. 10Qwen3.5 Omni Flash4.4 s
11. 11Amazon Nova Lite4.4 s
12. 12GPT-5 nano4.5 s
13. 13nemotron-3-nano-30b-a3b4.6 s
14. 14Mistral Small 44.6 s
15. 15Mistral Small (latest)4.7 s

SearchAll providersAlibabaAmazonAnt GroupAnt-groupAnthropicArcee AICohereDeepSeekGoogleIBMInceptionInclusionAIMetaMicrosoftMiniMaxMistral AIMoonshot AIMultiverse ComputingNous ResearchNVIDIAOpenAIStepFunThinking MachinesUpstagexAIXiaomiZhipu AIOpen weights426 of 426

#	Model	seconds ↓	Index ↕
1	GPT-5.4 ProOpenAI	10.1 min	64.2
2	GPT-5.4 ProOpenAI	10.1 min	61.1
3	GPT-6 AstraOpenAI	5.7 min	70.6
4	GPT-6 AstraOpenAI	5.7 min	69.9
5	GPT-6 AstraOpenAI	5.7 min	64.4
6	Claude Fable 5.1Anthropic	4.9 min	70.2
7	Claude Fable 5.1Anthropic	4.9 min	70.0
8	GPT-6 AstraOpenAI	3.7 min	68.6
9	GPT-5.6 TerraOpenAI	3.4 min	62.3
10	GPT-5.6 TerraOpenAI	3.4 min	58.0
11	Claude Sonnet 5Anthropic	3.3 min	61.2
12	Claude Sonnet 5Anthropic	3.3 min	50.7
13	GPT-5.6 LunaOpenAI	3.1 min	56.6
14	GPT-5.6 LunaOpenAI	3.1 min	56.5
15	GPT-5.6 SolOpenAI	2.5 min	65.5
16	GPT-5.6 SolOpenAI	2.5 min	65.3
17	GPT-5.2OpenAI	2.4 min	61.5
18	GPT-5.2OpenAI	2.4 min	58.5
19	GPT-5.2OpenAI	2.4 min	57.3
20	GPT-5.2OpenAI	2.4 min	55.5
21	GPT-5.2OpenAI	2.4 min	50.2
22	Claude Fable 5.1Anthropic	2.4 min	70.3
23	GPT-5.4 miniOpenAI	2.4 min	54.8
24	GPT-5.4 miniOpenAI	2.4 min	52.9
25	GPT-5.4 miniOpenAI	2.4 min	46.0
26	GPT-5.4OpenAI	2.2 min	63.4
27	GPT-5.4OpenAI	2.2 min	61.3
28	GPT-5.4OpenAI	2.2 min	58.8
29	GPT-5.4OpenAI	2.2 min	56.0
30	DeepSeek-R1-Distill-Llama-70BDeepSeekopen	2.1 min	42.2
31	GPT-5.4 nanoOpenAI	2.0 min	53.9
32	GPT-5.4 nanoOpenAI	2.0 min	48.3
33	GPT-5.4 nanoOpenAI	2.0 min	45.3
34	Gemma 3 12B ITNVIDIAopen	2.0 min	37.2
35	Qwen3.6 27BAlibabaopen	1.9 min	53.2
36	Qwen3.6 PlusAlibabaopen	1.9 min	59.0
37	GPT-6 AstraOpenAI	1.8 min	69.5
38	ByteDance-Seed/Seed-OSS-36B-InstructNVIDIA	1.7 min	48.0
39	Claude Fable 5Anthropic	1.7 min	70.4
40	Claude Fable 5Anthropic	1.7 min	65.8
41	Claude Fable 5Anthropic	1.7 min	63.9
42	Kimi K2.6Moonshot AIopen	1.7 min	60.6
43	GPT-5 nanoOpenAI	1.7 min	51.1
44	GPT-5 nanoOpenAI	1.7 min	45.9
45	GPT-5 nanoOpenAI	1.7 min	39.8
46	Qwen3.5-4BAlibaba	1.6 min	47.1
47	Claude Opus 5Anthropic	1.6 min	70.0
48	Claude Opus 5Anthropic	1.6 min	66.8
49	ALing Flash 2.0Ant-group	1.4 min	45.0
50	GPT-5OpenAI	1.4 min	59.5
51	GPT-5OpenAI	1.4 min	54.5
52	GPT-5 miniOpenAI	1.3 min	55.8
53	GPT-5 miniOpenAI	1.3 min	52.3
54	XMiMo-V2.5-ProXiaomiopen	1.3 min	59.0
55	GLM-5.1Zhipu AIopen	1.3 min	60.1
56	Kimi K3Moonshot AIopen	1.2 min	58.8
57	Kimi K3Moonshot AIopen	1.2 min	66.5
58	Kimi K3Moonshot AIopen	1.2 min	62.7
59	GLM-4.5VZhipu AIopen	1.2 min	40.6
60	Nemotron Nano 12B v2 VLNVIDIAopen	1.2 min	42.0

Show 100 more