Open-Weight LLM Leaderboard

Last updated Oct 29, 2026, 1:44 AM

Quick filters:

Showing 72 of 72 models

Ranking of 72 open-weight LLM models by Artificial Analysis intelligence, coding, and agentic benchmark scores, including context length and OpenRouter pricing.
#	Model				Context
1	Z.ai GLM 5.3	31.7	74.8	53.1	1.0M	$0.07 / $1.40
2	Xiaomi MiMo-V2.6-Pro	29.9	—	—	1.1M	$0.44 / $0.87
3	MiniMax MiniMax M3.1 Flash	28.9	—	—	1.0M	$0.15 / $0.60
4	StepFun Step 5 Preview	28.5	—	—	1.0M	$1.00 / $2.70
5	Xiaomi MiMo-V2.6-Flash	26.7	—	—	1.1M	$0.14 / $0.28
6	DeepSeek DeepSeek V4 Pro 0813	23.1	68.8	41.3	1.0M	$1.32 / $3.96
7	Qwen Qwen3.8 27B	20.1	68.1	45.8	1M	$0.04 / $1.25
8	MiniMax MiniMax M3	18.7	58.6	29.5	1.0M	$0.30 / $1.20
9	NVIDIA Nemotron 3.5 Super VL	13.6	—	—	1.0M	$0.15 / $0.50
10	Thinking Machines Inkling Small	12.9	52.9	23.5	524K	$0.45 / $1.20
11	Thinking Machines Inkling	12.5	52.1	22.5	524K	$1.00 / $4.05
12	NVIDIA Nemotron 3 Ultra	11.8	49.3	20.1	262K	$0.60 / $2.40
13	Meta Muse Glimmer 30B	6.3	49.0	8.5	131K	$0.30 / $1.20
14	inclusionAI Ling 3.0 Flash VL	—	57.0	28.7	131K	$0.06 / $0.18
15	IBM Granite 4.2 8B	—	22.4	1.3	131K	$0.06 / $0.15
16	Z.ai GLM 5.3 Flash	—	71.5	50.9	1.0M	$0.04 / $0.08
17	DeepSeek DeepSeek V4 Flash Vision Exp	—	65.0	47.5	1.0M	$0.22 / $0.65
18	Qwen Qwen3.8 2.4T A95B	—	71.9	50.1	1.0M	$2.00 / $6.00
19	LiquidAI LFM2.5-2.6B	—	7.7	—	128K	$0.35 / $0.50
20	NVIDIA Nemotron 3.5 Lightning	—	26.8	3.5	262K	$0.07 / $0.20
21	DeepSeek DeepSeek V4 Flash 0731	—	69.1	41.0	1.0M	$0.12 / $0.24
22	inclusionAI Ling 3.0 Flash	—	50.6	19.3	262K	$0.02 / $0.06
23	Meituan LongCat 2.0	—	45.3	14.0	1.0M	$0.30 / $1.20
24	MoonshotAI Kimi K3	—	76.2	50.0	1.0M	$0.19 / $20.00
25	Cohere North Mini Code	—	36.5	1.1	256K	$0.00 / $0.00
26	Z.ai GLM 5.2	—	68.8	38.4	1.0M	$0.01 / $7.00
27	MoonshotAI Kimi K2.7 Code	—	60.8	21.0	262K	$0.56 / $3.10
28	StepFun Step 3.7 Flash	—	39.6	—	262K	$0.20 / $1.15
29	NVIDIA Nemotron 3 Nano Omni	—	13.8	—	256K	$0.00 / $0.00
30	Qwen Qwen3.6 35B A3B	—	41.9	13.1	262K	$0.10 / $0.90
31	Qwen Qwen3.6 27B	—	53.7	18.5	262K	$0.45 / $2.70
32	DeepSeek DeepSeek V4 Pro 0423	—	59.4	26.3	1.0M	$0.96 / $1.91
33	DeepSeek DeepSeek V4 Flash 0423	—	56.2	22.2	1.0M	$0.09 / $0.18
34	Tencent Hy3 preview	—	58.8	24.1	262K	$0.18 / $0.60
35	Xiaomi MiMo-V2.5-Pro	—	60.2	21.3	1.1M	$0.48 / $1.80
36	Xiaomi MiMo-V2.5	—	56.8	15.8	1.1M	$0.40 / $2.00
37	MoonshotAI Kimi K2.6	—	61.8	20.5	262K	$0.65 / $3.41
38	Z.ai GLM 5.1	—	55.8	23.9	205K	$1.40 / $4.40
39	Google Gemma 4 26B A4B	—	39.3	2.9	262K	$0.08 / $0.26
40	Google Gemma 4 31B	—	43.4	4.2	262K	$0.20 / $0.40
41	MiniMax MiniMax M2.7	—	52.6	15.3	205K	$0.30 / $1.20
42	Mistral Mistral Small 4	—	26.6	0.8	262K	$0.15 / $0.60
43	NVIDIA Nemotron 3 Super	—	37.7	1.7	1.0M	$0.30 / $2.40
44	Qwen Qwen3.5-9B	—	28.7	1.2	262K	$0.17 / $0.25
45	Qwen Qwen3.5-35B-A3B	—	37.0	—	262K	$0.14 / $1.00
46	Qwen Qwen3.5-122B-A10B	—	45.7	7.6	262K	$0.26 / $2.08
47	Qwen Qwen3.5 397B A17B	—	48.2	8.3	262K	$0.50 / $3.60
48	Qwen Qwen3 Coder Next	—	36.2	0.9	262K	$0.12 / $0.80
49	MoonshotAI Kimi K2.5	—	46.8	—	262K	$0.50 / $2.50
50	Z.ai GLM 4.7	—	45.3	—	205K	$0.54 / $1.98
51	NVIDIA Nemotron 3 Nano 30B A3B	—	14.4	1.0	262K	$0.06 / $0.24
52	Mistral Devstral 2 2512	—	31.3	2.3	262K	$0.40 / $2.00
53	Mistral Ministral 3 14B 2512	—	14.4	1.1	262K	$0.20 / $0.20
54	Mistral Ministral 3 8B 2512	—	9.7	0.6	262K	$0.15 / $0.15
55	Mistral Ministral 3 3B 2512	—	4.8	0.8	131K	$0.10 / $0.10
56	DeepSeek DeepSeek V3.2	—	44.2	—	164K	$0.28 / $0.42
57	MoonshotAI Kimi K2 Thinking	—	21.0	—	262K	$0.60 / $2.50
58	Z.ai GLM 4.6	—	45.8	—	205K	$0.50 / $2.00
59	DeepSeek DeepSeek V3.1 Terminus	—	43.5	6.8	164K	$0.27 / $1.00
60	OpenAI gpt-oss-120b	—	30.4	3.7	131K	$0.15 / $0.60
61	OpenAI gpt-oss-20b	—	20.7	1.2	131K	$0.02 / $0.09
62	Qwen Qwen3 235B A22B Thinking 2507	—	22.1	1.3	128K	$0.45 / $3.50
63	Qwen Qwen3 14B	—	13.8	0.9	41K	$0.12 / $0.24
64	Qwen Qwen3 32B	—	15.3	0.9	131K	$0.08 / $0.28
65	Meta Llama 4 Maverick	—	16.3	0.6	524K	$0.19 / $0.65
66	Meta Llama 4 Scout	—	8.2	0.5	1.3M	$0.10 / $0.30
67	DeepSeek DeepSeek V3 0324	—	21.2	0.8	164K	$0.29 / $1.14
68	Google Gemma 3 4B	—	2.7	—	131K	$0.05 / $0.10
69	Google Gemma 3 12B	—	5.8	0.1	131K	$0.05 / $0.15
70	Google Gemma 3 27B	—	10.1	0.1	131K	$0.08 / $0.35
71	Meta Llama 3.3 70B Instruct	—	11.9	—	131K	$0.10 / $0.32
72	Meta Llama 3.1 8B Instruct	—	5.4	—	131K	$0.05 / $0.08

Z.ai
GLM 5.3
Intel.
31.7
Coding
74.8
Agentic
53.1
1.0M ctx$0.07 / $1.40 /1M
Xiaomi
MiMo-V2.6-Pro
Intel.
29.9
Coding
—
Agentic
—
1.1M ctx$0.44 / $0.87 /1M
MiniMax
MiniMax M3.1 Flash
Intel.
28.9
Coding
—
Agentic
—
1.0M ctx$0.15 / $0.60 /1M
StepFun
Step 5 Preview
Intel.
28.5
Coding
—
Agentic
—
1.0M ctx$1.00 / $2.70 /1M
Xiaomi
MiMo-V2.6-Flash
Intel.
26.7
Coding
—
Agentic
—
1.1M ctx$0.14 / $0.28 /1M
DeepSeek
DeepSeek V4 Pro 0813
Intel.
23.1
Coding
68.8
Agentic
41.3
1.0M ctx$1.32 / $3.96 /1M
Qwen
Qwen3.8 27B
Intel.
20.1
Coding
68.1
Agentic
45.8
1M ctx$0.04 / $1.25 /1M
MiniMax
MiniMax M3
Intel.
18.7
Coding
58.6
Agentic
29.5
1.0M ctx$0.30 / $1.20 /1M
NVIDIA
Nemotron 3.5 Super VL
Intel.
13.6
Coding
—
Agentic
—
1.0M ctx$0.15 / $0.50 /1M
Thinking Machines
Inkling Small
Intel.
12.9
Coding
52.9
Agentic
23.5
524K ctx$0.45 / $1.20 /1M
Thinking Machines
Inkling
Intel.
12.5
Coding
52.1
Agentic
22.5
524K ctx$1.00 / $4.05 /1M
NVIDIA
Nemotron 3 Ultra
Intel.
11.8
Coding
49.3
Agentic
20.1
262K ctx$0.60 / $2.40 /1M
Meta
Muse Glimmer 30B
Intel.
6.3
Coding
49.0
Agentic
8.5
131K ctx$0.30 / $1.20 /1M
inclusionAI
Ling 3.0 Flash VL
Intel.
—
Coding
57.0
Agentic
28.7
131K ctx$0.06 / $0.18 /1M
IBM
Granite 4.2 8B
Intel.
—
Coding
22.4
Agentic
1.3
131K ctx$0.06 / $0.15 /1M
Z.ai
GLM 5.3 Flash
Intel.
—
Coding
71.5
Agentic
50.9
1.0M ctx$0.04 / $0.08 /1M
DeepSeek
DeepSeek V4 Flash Vision Exp
Intel.
—
Coding
65.0
Agentic
47.5
1.0M ctx$0.22 / $0.65 /1M
Qwen
Qwen3.8 2.4T A95B
Intel.
—
Coding
71.9
Agentic
50.1
1.0M ctx$2.00 / $6.00 /1M
LiquidAI
LFM2.5-2.6B
Intel.
—
Coding
7.7
Agentic
—
128K ctx$0.35 / $0.50 /1M
NVIDIA
Nemotron 3.5 Lightning
Intel.
—
Coding
26.8
Agentic
3.5
262K ctx$0.07 / $0.20 /1M
DeepSeek
DeepSeek V4 Flash 0731
Intel.
—
Coding
69.1
Agentic
41.0
1.0M ctx$0.12 / $0.24 /1M
inclusionAI
Ling 3.0 Flash
Intel.
—
Coding
50.6
Agentic
19.3
262K ctx$0.02 / $0.06 /1M
Meituan
LongCat 2.0
Intel.
—
Coding
45.3
Agentic
14.0
1.0M ctx$0.30 / $1.20 /1M
MoonshotAI
Kimi K3
Intel.
—
Coding
76.2
Agentic
50.0
1.0M ctx$0.19 / $20.00 /1M
Cohere
North Mini Code
Intel.
—
Coding
36.5
Agentic
1.1
256K ctx$0.00 / $0.00 /1M
Z.ai
GLM 5.2
Intel.
—
Coding
68.8
Agentic
38.4
1.0M ctx$0.01 / $7.00 /1M
MoonshotAI
Kimi K2.7 Code
Intel.
—
Coding
60.8
Agentic
21.0
262K ctx$0.56 / $3.10 /1M
StepFun
Step 3.7 Flash
Intel.
—
Coding
39.6
Agentic
—
262K ctx$0.20 / $1.15 /1M
NVIDIA
Nemotron 3 Nano Omni
Intel.
—
Coding
13.8
Agentic
—
256K ctx$0.00 / $0.00 /1M
Qwen
Qwen3.6 35B A3B
Intel.
—
Coding
41.9
Agentic
13.1
262K ctx$0.10 / $0.90 /1M
Qwen
Qwen3.6 27B
Intel.
—
Coding
53.7
Agentic
18.5
262K ctx$0.45 / $2.70 /1M
DeepSeek
DeepSeek V4 Pro 0423
Intel.
—
Coding
59.4
Agentic
26.3
1.0M ctx$0.96 / $1.91 /1M
DeepSeek
DeepSeek V4 Flash 0423
Intel.
—
Coding
56.2
Agentic
22.2
1.0M ctx$0.09 / $0.18 /1M
Tencent
Hy3 preview
Intel.
—
Coding
58.8
Agentic
24.1
262K ctx$0.18 / $0.60 /1M
Xiaomi
MiMo-V2.5-Pro
Intel.
—
Coding
60.2
Agentic
21.3
1.1M ctx$0.48 / $1.80 /1M
Xiaomi
MiMo-V2.5
Intel.
—
Coding
56.8
Agentic
15.8
1.1M ctx$0.40 / $2.00 /1M
MoonshotAI
Kimi K2.6
Intel.
—
Coding
61.8
Agentic
20.5
262K ctx$0.65 / $3.41 /1M
Z.ai
GLM 5.1
Intel.
—
Coding
55.8
Agentic
23.9
205K ctx$1.40 / $4.40 /1M
Google
Gemma 4 26B A4B
Intel.
—
Coding
39.3
Agentic
2.9
262K ctx$0.08 / $0.26 /1M
Google
Gemma 4 31B
Intel.
—
Coding
43.4
Agentic
4.2
262K ctx$0.20 / $0.40 /1M
MiniMax
MiniMax M2.7
Intel.
—
Coding
52.6
Agentic
15.3
205K ctx$0.30 / $1.20 /1M
Mistral
Mistral Small 4
Intel.
—
Coding
26.6
Agentic
0.8
262K ctx$0.15 / $0.60 /1M
NVIDIA
Nemotron 3 Super
Intel.
—
Coding
37.7
Agentic
1.7
1.0M ctx$0.30 / $2.40 /1M
Qwen
Qwen3.5-9B
Intel.
—
Coding
28.7
Agentic
1.2
262K ctx$0.17 / $0.25 /1M
Qwen
Qwen3.5-35B-A3B
Intel.
—
Coding
37.0
Agentic
—
262K ctx$0.14 / $1.00 /1M
Qwen
Qwen3.5-122B-A10B
Intel.
—
Coding
45.7
Agentic
7.6
262K ctx$0.26 / $2.08 /1M
Qwen
Qwen3.5 397B A17B
Intel.
—
Coding
48.2
Agentic
8.3
262K ctx$0.50 / $3.60 /1M
Qwen
Qwen3 Coder Next
Intel.
—
Coding
36.2
Agentic
0.9
262K ctx$0.12 / $0.80 /1M
MoonshotAI
Kimi K2.5
Intel.
—
Coding
46.8
Agentic
—
262K ctx$0.50 / $2.50 /1M
Z.ai
GLM 4.7
Intel.
—
Coding
45.3
Agentic
—
205K ctx$0.54 / $1.98 /1M
NVIDIA
Nemotron 3 Nano 30B A3B
Intel.
—
Coding
14.4
Agentic
1.0
262K ctx$0.06 / $0.24 /1M
Mistral
Devstral 2 2512
Intel.
—
Coding
31.3
Agentic
2.3
262K ctx$0.40 / $2.00 /1M
Mistral
Ministral 3 14B 2512
Intel.
—
Coding
14.4
Agentic
1.1
262K ctx$0.20 / $0.20 /1M
Mistral
Ministral 3 8B 2512
Intel.
—
Coding
9.7
Agentic
0.6
262K ctx$0.15 / $0.15 /1M
Mistral
Ministral 3 3B 2512
Intel.
—
Coding
4.8
Agentic
0.8
131K ctx$0.10 / $0.10 /1M
DeepSeek
DeepSeek V3.2
Intel.
—
Coding
44.2
Agentic
—
164K ctx$0.28 / $0.42 /1M
MoonshotAI
Kimi K2 Thinking
Intel.
—
Coding
21.0
Agentic
—
262K ctx$0.60 / $2.50 /1M
Z.ai
GLM 4.6
Intel.
—
Coding
45.8
Agentic
—
205K ctx$0.50 / $2.00 /1M
DeepSeek
DeepSeek V3.1 Terminus
Intel.
—
Coding
43.5
Agentic
6.8
164K ctx$0.27 / $1.00 /1M
OpenAI
gpt-oss-120b
Intel.
—
Coding
30.4
Agentic
3.7
131K ctx$0.15 / $0.60 /1M
OpenAI
gpt-oss-20b
Intel.
—
Coding
20.7
Agentic
1.2
131K ctx$0.02 / $0.09 /1M
Qwen
Qwen3 235B A22B Thinking 2507
Intel.
—
Coding
22.1
Agentic
1.3
128K ctx$0.45 / $3.50 /1M
Qwen
Qwen3 14B
Intel.
—
Coding
13.8
Agentic
0.9
41K ctx$0.12 / $0.24 /1M
Qwen
Qwen3 32B
Intel.
—
Coding
15.3
Agentic
0.9
131K ctx$0.08 / $0.28 /1M
Meta
Llama 4 Maverick
Intel.
—
Coding
16.3
Agentic
0.6
524K ctx$0.19 / $0.65 /1M
Meta
Llama 4 Scout
Intel.
—
Coding
8.2
Agentic
0.5
1.3M ctx$0.10 / $0.30 /1M
DeepSeek
DeepSeek V3 0324
Intel.
—
Coding
21.2
Agentic
0.8
164K ctx$0.29 / $1.14 /1M
Google
Gemma 3 4B
Intel.
—
Coding
2.7
Agentic
—
131K ctx$0.05 / $0.10 /1M
Google
Gemma 3 12B
Intel.
—
Coding
5.8
Agentic
0.1
131K ctx$0.05 / $0.15 /1M
Google
Gemma 3 27B
Intel.
—
Coding
10.1
Agentic
0.1
131K ctx$0.08 / $0.35 /1M
Meta
Llama 3.3 70B Instruct
Intel.
—
Coding
11.9
Agentic
—
131K ctx$0.10 / $0.32 /1M
Meta
Llama 3.1 8B Instruct
Intel.
—
Coding
5.4
Agentic
—
131K ctx$0.05 / $0.08 /1M

Data is sourced live from OpenRouter's public model catalog and Artificial Analysis benchmark scores, filtered to models with public Hugging Face weights, and refreshed automatically throughout the day. Scores are on Artificial Analysis's 0–100 scale. LMSA does not represent the benchmark results shown on this page.