Leaderboard | Text

Models

Choose model family

Claude by Anthropic

Mistral by Mistral AI

Choose topic

All topics Facts and Information Creative Writing and Ideation Logic and Problem-Solving Task Completion Coding

Language

Choose language

All languages English Chinese Arabic Spanish Indonesian Japanese

More filters

Show inactive models

Hide models that are no longer actively available on Yupp.

Turns

Filter model performance by the number of turns in a conversation.

All Single turn Multiple turns

Open license models

Filter the leaderboard to only show models that have an open license.

All selected Open license Proprietary license

546

Qwen 2.5 VL 3B Instruct

726

GLM 4.6V Flash

744

Pixtral 12B

778

Pixtral Large

821

Nemotron 3 Nano (Thinking)

853

Qwen3 32B Fast

856

Qwen3 14B

869

DeepSeek-R1 Distill Llama 70B

883

Llama 4 Maverick

894

DeepSeek-R1

898

Qwen3 235B A22B

910

QwQ 32B

918

Qwen3 30B A3B

947

Mistral Large 3

950

gpt-oss-20b

Last updated about 1 month ago

Rank	Overall	Name	VIBE Score	Confidence Interval	Votes	Downvote %	Abort %	Speed	Latency	Context	Cost (Input)	Cost (Output)
1	288	Qwen 2.5 VL 3B Instruct	546	±31	1K	9.1%	3.0%	44 tps	2.5s	128K	$0.21	$0.63
2	186	GLM 4.6V Flash	726	±29	575	2.5%	3.7%	64 tps	2.1s	128K	$0.04	$0.40
3	274	Pixtral 12B	744	±33	940	9.6%	2.2%	101 tps	1.2s	131K	$0.08	$0.08
4	165	Pixtral Large	778	±18	1.1K	7.2%	2.5%	57 tps	1.3s	128K	$1.50	$4.50
5	86	Nemotron 3 Nano (Thinking)	821	±23	540	4.4%	2.0%	200 tps	0.5s	256K	$0	$0
6	121	Qwen3 32B Fast	853	±13	1.8K	4.2%	11.6%	30 tps	3.1s	41K	$0.10	$0.25
7	133	Qwen3 14B	856	±24	745	2.6%	1.7%	109 tps	0.8s	41K	$0.04	$0.15
8	246	DeepSeek-R1 Distill Llama 70B	869	±28	590	4.8%	3.6%	27 tps	1.6s	32K	$0.73	$0.95
9	161	Llama 4 Maverick	883	±12	3.6K	4.4%	1.2%	88 tps	2.4s	1M	$0.23	$0.83
10	148	DeepSeek-R1	894	±12	1.1K	3.8%	0.8%	133 tps	0.6s	64K	$0.91	$3.07
11	86	Qwen3 235B A22B	898	±23	725	3.3%	5.3%	71 tps	0.9s	41K	$0.23	$0.63
12	121	QwQ 32B	910	±13	1.8K	3.1%	5.4%	41 tps	2.1s	16K	$0.43	$0.56
13	126	Qwen3 30B A3B	918	±15	865	3.4%	5.1%	163 tps	1.0s	41K	$0.06	$0.21
14	65	Mistral Large 3	947	±20	1.3K	4.4%	2.1%	51 tps	1.0s	256K	$0.50	$1.50
15	101	gpt-oss-20b	950	±18	1.4K	4.7%	0.5%	216 tps	0.5s	131K	$0.06	$0.26
16	126	DeepSeek V3	960	±7	3.4K	2.3%	0.9%	69 tps	1.1s	64K	$0.59	$1.49
17	129	Command A	965	±8	3K	2.9%	2.2%	42 tps	0.8s	256K	$2.00	$7.33
18	113	Kimi K2 Fast	975	±10	4.8K	2.3%	0.8%	365 tps	0.5s	131K	$1.00	$3.00
19	106	DeepSeek V3.1 Terminus Thinking	1000	±14	745	2.6%	5.9%	27 tps	1.8s	131K	$0.56	$1.68
20	133	DeepSeek-R1 0528	1001	±15	1.1K	4.1%	1.3%	93 tps	0.5s	64K	$1.60	$3.67
21	95	DeepSeek-R1 Turbo	1009	±20	660	3.6%	2.6%	29 tps	1.8s	64K	$2.85	$4.75
22	56	DeepSeek V3.2 Thinking	1021	±13	1.9K	1.8%	9.0%	30 tps	2.6s	131K	$0.28	$0.42
23	79	MiniMax M2.5 Lightning	1031	±20	820	1.8%	1.5%	51 tps	2.0s	205K	$0.60	$2.40
24	95	DeepSeek V3.2 Exp Thinking	1038	±17	655	3.7%	7.2%	26 tps	3.0s	131K	$0.28	$0.42
25	81	Qwen3.5 27B	1056	±17	665	2.9%	3.7%	55 tps	2.6s	256K	$0.30	$2.40
26	52	Qwen3.5 122B A17B	1063	±14	980	1.5%	1.5%	82 tps	1.4s	256K	$0.40	$3.20
27	48	Step 3.5 Flash	1067	±20	630	2.3%	2.2%	109 tps	0.6s	256K	$0.05	$0.15
28	65	DeepSeek V3.2 Exp Chat	1072	±12	755	2.6%	2.6%	29 tps	1.5s	131K	$0.27	$0.39
29	44	Kimi K2 Thinking Turbo	1072	±17	1.3K	2.2%	2.0%	75 tps	1.4s	262K	$1.15	$8.00
30	48	gpt-oss-120b	1074	±6	3K	2.6%	0.7%	213 tps	0.5s	131K	$0.11	$0.50
31	33	Qwen3 Next 80B A3B Instruct	1083	±16	1.5K	2.6%	0.6%	84 tps	1.1s	256K	$0.20	$1.42
32	33	Kimi K2.5	1090	±13	4.3K	2.1%	6.5%	33 tps	1.7s	262K	$0.34	$2.57
33	37	Kimi K2.5 Instant	1093	±12	1.4K	2.7%	2.9%	32 tps	3.0s	262K	$0.50	$3.00