飞哩飞哩

leaderboard AI 基准评测

来自 Artificial Analysis 的综合模型能力评测数据，实时同步。

AA Agentic Index· agent

518 个模型 · 满分 1.0

GPT-5.5 (xhigh)
74.100
GPT-5.5 (high)
72.000
Claude Opus 4.7 (Adaptive Reasoning, Max Effort)
71.300
Gemini 3.5 Flash (high)
70.300
GPT-5.5 (medium)
69.400
GPT-5.4 (xhigh)
68.000
Claude Opus 4.6 (Adaptive Reasoning, Max Effort)
67.600
MiMo-V2.5-Pro
67.400
DeepSeek V4 Pro (Reasoning, Max Effort)
67.200
GLM-5.1 (Reasoning)
67.100
DeepSeek V4 Pro (Reasoning, High Effort)
66.700
Qwen3.7 Max
66.600
GLM-5-Turbo
66.100
Kimi K2.6
66.000
GLM-5.1 (Non-reasoning)
66.000
Grok 4.3 (high)
65.900
MiMo-V2.5
65.500
Qwen3.6 Max Preview
64.800
Claude Opus 4.7 (Non-reasoning, High Effort)
64.600
Claude Opus 4.6 (Non-reasoning, High Effort)
64.200
DeepSeek V4 Pro (Non-reasoning)
63.300
GLM-5 (Reasoning)
63.100
Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort)
63.000
Qwen3.6 27B (Reasoning)
62.900
MiMo-V2-Pro
62.800
DeepSeek V4 Flash (Reasoning, High Effort)
62.300
Muse Spark
62.000
Qwen3.6 Plus
61.700
Claude Sonnet 4.6 (Non-reasoning, High Effort)
61.600
MiniMax-M2.7
61.500
DeepSeek V4 Flash (Non-reasoning)
61.300
DeepSeek V4 Flash (Reasoning, Max Effort)
61.300
GLM 5V Turbo (Reasoning)
61.100
Qwen3.6 27B (Non-reasoning)
60.900
GPT-5.3 Codex (xhigh)
60.500
GLM-5 (Non-reasoning)
60.300
GPT-5.2 (xhigh)
60.200
GPT-5.5 (low)
59.700
Claude Opus 4.5 (Reasoning)
59.600
Claude Opus 4.5 (Non-reasoning)
59.200
Grok-1
59.100
GPT-5.5 Pro (xhigh)
59.100
Llama 3.2 Instruct 90B (Vision)
59.100
Gemini 3.1 Pro Preview
59.100
DeepSeek R1 Distill Llama 70B
59.100
R1 1776
59.100
Phi-4 Multimodal Instruct
59.100
Llama 3.3 Nemotron Super 49B v1 (Reasoning)
59.100
Llama 3.3 Nemotron Super 49B v1 (Non-reasoning)
59.100
Llama 3.1 Nemotron Nano 4B v1.1 (Reasoning)
59.100
Llama 65B
59.100
Kimi Linear 48B A3B Instruct
59.100
Molmo 7B-D
59.100
DeepHermes 3 - Mistral 24B Preview (Non-reasoning)
59.100
DeepHermes 3 - Llama-3.1 8B Preview (Non-reasoning)
59.100
EXAONE 4.5 33B (Non-reasoning)
59.100
Qwen Chat 14B
59.100
Cogito v2.1 (Reasoning)
59.100
Mi:dm K 2.5 Pro Preview
59.100
o1-preview
59.100
o1-mini
59.100
GPT-4o (May '24)
59.100
GPT-4 Turbo
59.100
GPT-4o mini
59.100
GPT-3.5 Turbo
59.100
GPT-4o (ChatGPT)
59.100
o1-pro
59.100
GPT-4o Realtime (Dec '24)
59.100
GPT-5 (ChatGPT)
59.100
GPT-4o mini Realtime (Dec '24)
59.100
o3-pro
59.100
GPT-4o (March 2025, chatgpt-4o-latest)
59.100
GPT-4
59.100
GPT-4.5 (Preview)
59.100
o3-mini
59.100
GPT-5.4 Pro (xhigh)
59.100
GPT-3.5 Turbo (0613)
59.100
Llama 3.2 Instruct 3B
59.100
Llama 2 Chat 7B
59.100
Llama 2 Chat 70B
59.100
Llama 2 Chat 13B
59.100
Gemini 2.0 Pro Experimental (Feb '25)
59.100
Gemini 2.0 Flash (experimental)
59.100
Gemini 1.5 Pro (Sep '24)
59.100
Gemini 2.0 Flash-Lite (Preview)
59.100
Gemini 1.5 Flash (Sep '24)
59.100
Gemini 1.5 Flash-8B
59.100
Gemini 1.0 Pro
59.100
Gemini 2.0 Flash-Lite (Feb '25)
59.100
Gemini 2.5 Flash Preview (Non-reasoning)
59.100
Gemini 1.5 Flash (May '24)
59.100
Gemini 2.5 Flash Preview (Reasoning)
59.100
PALM-2
59.100
Gemini 1.5 Pro (May '24)
59.100
Gemini 2.0 Flash Thinking Experimental (Jan '25)
59.100
Gemini 2.5 Pro Preview (Mar' 25)
59.100
Gemini 1.0 Ultra
59.100
Gemma 3n E4B Instruct Preview (May '25)
59.100
Gemini 2.0 Flash Thinking Experimental (Dec '24)
59.100
Gemini 2.5 Pro Preview (May' 25)
59.100
Claude 3.5 Sonnet (Oct '24)
59.100
Claude 3.5 Sonnet (June '24)
59.100
Claude 3 Opus
59.100
Claude 3 Sonnet
59.100
Claude Instant
59.100
Claude 4.1 Opus (Non-reasoning)
59.100
Claude 2.0
59.100
Claude 4 Opus (Non-reasoning)
59.100
Claude 4.1 Opus (Reasoning)
59.100
Claude 4 Opus (Reasoning)
59.100
Claude 2.1
59.100
Mistral Large 2 (Jul '24)
59.100
Pixtral Large
59.100
Mistral Small 3
59.100
Mistral Small (Sep '24)
59.100
Mixtral 8x22B Instruct
59.100
Mistral Small (Feb '24)
59.100
Mistral Large (Feb '24)
59.100
Mixtral 8x7B Instruct
59.100
Mistral 7B Instruct
59.100
Magistral Small 1
59.100
Mistral Medium
59.100
Mistral Saba
59.100
DeepSeek R1 Distill Qwen 32B
59.100
DeepSeek R1 Distill Qwen 14B
59.100
DeepSeek-V2.5 (Dec '24)
59.100
DeepSeek-Coder-V2
59.100
DeepSeek R1 Distill Llama 8B
59.100
DeepSeek LLM 67B Chat (V1)
59.100
DeepSeek R1 Distill Qwen 1.5B
59.100
DeepSeek-V2-Chat
59.100
DeepSeek Coder V2 Lite Instruct
59.100
DeepSeek-V2.5
59.100
DeepSeek R1 0528 Qwen3 8B
59.100
Sonar
59.100
Sonar Pro
59.100
Sonar Reasoning Pro
59.100
Sonar Reasoning
59.100
Grok Beta
59.100
Grok 3 Reasoning Beta
59.100
Grok 2 (Dec '24)
59.100
OpenChat 3.5 (1210)
59.100
Phi-3 Mini Instruct 3.8B
59.100
LFM 40B
59.100
Solar Mini
59.100
Solar Pro 2 (Preview) (Reasoning)
59.100
Solar Pro 2 (Preview) (Non-reasoning)
59.100
DBRX Instruct
59.100
MiniMax M1 40k
59.100
Llama 3.1 Tulu3 405B
59.100
OLMo 2 32B
59.100
OLMo 2 7B
59.100
Reka Flash (Sep '24)
59.100
Hermes 3 - Llama-3.1 70B
59.100
Command-R+ (Apr '24)
59.100
Command-R (Mar '24)
59.100
Apriel-v1.5-15B-Thinker
59.100
Jamba 1.5 Large
59.100
Jamba 1.6 Large
59.100
Jamba 1.5 Mini
59.100
Jamba 1.6 Mini
59.100
Arctic Instruct
59.100
Qwen2.5 Max
59.100
Qwen2.5 Instruct 72B
59.100
Qwen2.5 Coder Instruct 32B
59.100
Qwen2.5 Turbo
59.100
Qwen2 Instruct 72B
59.100
Qwen3 32B (Non-reasoning)
59.100
QwQ 32B
59.100
Qwen3 4B (Non-reasoning)
59.100
GPT-5.4 mini (xhigh)
58.900
Kimi K2.5 (Reasoning)
58.900
Kimi K2.6 (Non-reasoning)
58.700
MiMo-V2-Omni
58.600
MiMo-V2-Omni-0327
58.600
Qwen3.6 35B A3B (Reasoning)
58.300
GPT-5.4 (low)
58.200
Claude Sonnet 4.6 (Non-reasoning, Low Effort)
57.500
GPT-5.2 Codex (xhigh)
56.500
Qwen3.5 397B A17B (Reasoning)
55.800
Hy3-preview (Reasoning)
55.700
MiniMax-M2.5
55.600
GLM-4.7 (Reasoning)
55.000
GPT-5.2 (medium)
54.900
GPT-5 (high)
54.700
Qwen3.5 27B (Reasoning)
54.600
GLM-4.7 (Non-reasoning)
54.300
Grok 4.20 0309 v2 (Reasoning)
53.900
Qwen3.5 397B A17B (Non-reasoning)
53.300
Mistral Medium 3.5
53.200
Qwen3.5 122B A10B (Reasoning)
53.000
DeepSeek V3.2 (Reasoning)
52.900
Kimi K2.5 (Non-reasoning)
52.800
Qwen3.5 Omni Plus
52.800
GPT-5 Codex (high)
52.700
Qwen3.6 35B A3B (Non-reasoning)
52.500
JT-35B-Flash
52.300
MiMo-V2-Flash (Reasoning)
52.100
Gemini 3 Pro Preview (high)
52.000
Step 3.5 Flash
52.000
Claude 4.5 Sonnet (Reasoning)
51.700
Ring-2.6-1T
51.500
Qwen3.5 27B (Non-reasoning)
51.500
GPT-5.1 (high)
51.300
Grok 4.20 0309 (Reasoning)
50.900
MiMo-V2.5-Pro (Non-reasoning)
50.800
KAT Coder Pro V2
50.700
GPT-5.1 Codex (high)
50.700
Claude 4.5 Sonnet (Non-reasoning)
50.600
GPT-5.5 (Non-reasoning)
50.200
Qwen3 Max Thinking
50.100
Gemini 3 Flash Preview (Reasoning)
49.700
GPT-5 (low)
49.700
Qwen3.5 122B A10B (Non-reasoning)
49.500
Grok 4.1 Fast (Reasoning)
49.300
Grok 4.3 (Non-reasoning)
48.800
MiMo-V2-Flash (Feb 2026)
48.800
Step 3.5 Flash 2603
48.200
Ling-2.6-1T
48.200
Qwen3.5 35B A3B (Non-reasoning)
48.000
Kimi K2 Thinking
47.900
GPT-5.4 nano (xhigh)
47.600
MiniMax-M2
47.500
MiniMax-M2.1
47.400
MiMo-V2-Flash (Non-reasoning)
47.300
Nova 2.0 Pro Preview (medium)
47.000
Hy3-preview (Non-reasoning)
46.700
GLM-4.7-Flash (Reasoning)
46.000
GPT-5 (medium)
45.800
GPT-5 mini (high)
45.500
Gemini 3 Pro Preview (low)
45.000
Qwen3.5 35B A3B (Reasoning)
44.100
Claude 4 Sonnet (Reasoning)
43.000
Qwen3 Max
43.000
Qwen3 Max Thinking (Preview)
43.000
GLM-4.6 (Non-reasoning)
42.900
Trinity Large Thinking
42.600
JT-MINI
42.400
Qwen3 Coder Next
42.100
KAT-Coder-Pro V1
41.700
GPT-5.4 nano (medium)
41.600
Qwen3.5 Omni Flash
41.600
GLM-4.6 (Reasoning)
41.600
Grok 4
41.500
GLM-4.7-Flash (Non-reasoning)
41.400
Qwen3.5 9B (Non-reasoning)
41.100
Gemma 4 31B (Reasoning)
40.900
GPT-5 mini (medium)
40.900
GPT-5.4 mini (medium)
40.300
Claude 4.5 Haiku (Reasoning)
40.200
NVIDIA Nemotron 3 Super 120B A12B (Reasoning)
40.200
DeepSeek V3.2 (Non-reasoning)
39.800
Mercury 2
39.700
ERNIE 5.0 Thinking Preview
39.700
GPT-5.2 (Non-reasoning)
39.500
Grok 4 Fast (Reasoning)
39.500
Gemma 4 31B (Non-reasoning)
39.400
Claude 4 Sonnet (Non-reasoning)
39.200
GPT-5.4 (Non-reasoning)
39.100
LongCat Flash Lite
38.800
GPT-5.1 Codex mini (high)
38.700
Grok 4.20 0309 v2 (Non-reasoning)
38.300
Nova 2.0 Omni (medium)
38.200
K-EXAONE (Reasoning)
38.100
Ling 2.6 Flash
38.100
HyperCLOVA X SEED Think (32B)
38.000
gpt-oss-120b (high)
37.900
Grok 4.20 0309 (Non-reasoning)
37.800
Nova 2.0 Pro Preview (low)
37.700
Kimi K2 0905
37.700
Qwen3.5 9B (Reasoning)
37.400
Nova 2.0 Lite (high)
37.300
Claude 3.7 Sonnet (Reasoning)
37.000
Mi:dm K 2.5 Pro
36.800
EXAONE 4.5 33B
36.500
Doubao Seed Code
36.400
Qwen3.5 4B (Non-reasoning)
36.300
o3
36.100
o4-mini (high)
36.100
Claude 3.7 Sonnet (Non-reasoning)
35.700
Grok Code Fast 1
35.600
Gemini 3 Flash Preview (Non-reasoning)
35.000
Solar Pro 3
34.900
Gemini 2.5 Flash Preview (Sep '25) (Reasoning)
34.700
Grok 4.1 Fast (Non-reasoning)
32.900
Nova 2.0 Lite (medium)
32.900
Gemini 2.5 Pro
32.700
Claude 4.5 Haiku (Non-reasoning)
32.600
Qwen3.5 4B (Reasoning)
32.500
GPT-5.1 (Non-reasoning)
32.200
Gemma 4 26B A4B (Reasoning)
32.100
Grok 4 Fast (Non-reasoning)
32.000
DeepSeek V3.1 (Non-reasoning)
31.900
Grok 3 mini Reasoning (high)
31.200
K-EXAONE (Non-reasoning)
31.200
Tri-21B-think Preview
31.100
o1
31.100
DeepSeek V3.2 Exp (Non-reasoning)
31.000
DeepSeek V3.1 Terminus (Reasoning)
29.800
Qwen3 235B A22B 2507 (Reasoning)
29.700
MiniCPM-V 4.6 1.3B
29.200
MiniMax M1 80k
29.100
Gemma 4 26B A4B (Non-reasoning)
28.900
DeepSeek V3.2 Exp (Reasoning)
28.700
Apriel-v1.6-15B-Thinker
28.600
DeepSeek V3.1 Terminus (Non-reasoning)
28.600
gpt-oss-120b (low)
28.000
Nova 2.0 Lite (low)
27.800
Seed-OSS-36B-Instruct
27.700
gpt-oss-20B (high)
27.600
GPT-4.1
27.300
Qwen3.5 2B (Non-reasoning)
27.200
Tri-21B-Think
27.000
Qwen3 VL 235B A22B (Reasoning)
27.000
Nemotron Cascade 2 30B A3B
26.200
GPT-5.4 nano (Non-Reasoning)
25.900
Mistral Small 4 (Reasoning)
25.900
Gemini 3.1 Flash-Lite Preview
25.700
Mistral Medium 3.1
25.300
GPT-4.1 mini
25.200
Devstral Small (May '25)
25.100
GPT-5.4 mini (Non-Reasoning)
25.000
Sarvam 105B (high)
24.700
Grok 3
24.600
Magistral Medium 1.2
24.500
Solar Open 100B (Reasoning)
24.300
Kimi K2
24.300
Nova 2.0 Pro Preview (Non-reasoning)
23.900
Nemotron 3 Nano Omni 30B A3B Reasoning
23.900
Qwen3 Next 80B A3B (Reasoning)
23.600
Qwen3 VL 32B (Reasoning)
23.400
Qwen3 Max (Preview)
23.300
GPT-5 nano (high)
23.200
Qwen3.5 2B (Reasoning)
23.000
Gemini 2.5 Flash Preview (Sep '25) (Non-reasoning)
23.000
GPT-5 (minimal)
22.900
Qwen3 235B A22B 2507 Instruct
22.800
Nova 2.0 Omni (low)
22.600
gpt-oss-20B (low)
21.900
Devstral 2
21.900
Qwen2.5 Instruct 32B
21.900
Qwen1.5 Chat 110B
21.900
Qwen Chat 72B
21.900
Qwen3 4B (Reasoning)
21.900
Qwen2.5 Coder Instruct 7B
21.900
Mistral Large 3
21.700
Qwen3.5 0.8B (Non-reasoning)
21.700
Nova 2.0 Lite (Non-reasoning)
21.100
Qwen3 Coder 30B A3B Instruct
21.100
GLM-4.5-Air
21.000
o3-mini (high)
20.900
Devstral Small 2
20.800
DeepSeek R1 0528 (May '25)
20.800
INTELLECT-3
19.800
Motif-2-12.7B-Reasoning
19.200
Qwen3 235B A22B (Non-reasoning)
19.200
NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)
19.100
Qwen3 VL 235B A22B Instruct
19.100
DeepSeek V3.1 (Reasoning)
18.900
Gemini 2.5 Flash (Reasoning)
18.800
GLM-4.6V (Non-reasoning)
18.700
Mistral Small 4 (Non-reasoning)
18.600
Qwen3 235B A22B (Reasoning)
18.400
Qwen3 VL 8B Instruct
18.400
Ring-1T
18.300
Qwen3 Coder 480B A35B Instruct
18.300
Claude 3.5 Haiku
17.700
Qwen3 30B A3B 2507 (Reasoning)
17.700
GLM-4.6V (Reasoning)
17.500
Ministral 3 14B
17.400
Magistral Small 1.2
17.300
GPT-5 nano (medium)
16.800
Ministral 3 8B
16.700
Nova Premier
16.400
DeepSeek V3 0324
16.300
GLM-4.5 (Reasoning)
16.200
Qwen3.5 0.8B (Reasoning)
15.900
Gemini 2.0 Flash (Feb '25)
15.800
Devstral Medium
15.800
Magistral Medium 1
15.700
Qwen3 VL 8B (Reasoning)
15.600
K2 Think V2
15.300
Gemini 2.5 Flash (Non-reasoning)
15.000
Nova 2.0 Omni (Non-reasoning)
14.900
Qwen3 VL 30B A3B (Reasoning)
14.900
Qwen3 4B 2507 (Reasoning)
14.600
K2-V2 (high)
14.400
Qwen3 VL 4B (Reasoning)
14.400
Qwen3 14B (Reasoning)
14.400
Devstral Small (Jul '25)
14.300
Qwen3 Next 80B A3B Instruct
14.200
Granite 4.1 30B
14.000
K2-V2 (medium)
13.900
Mistral Medium 3
13.700
Qwen3 14B (Non-reasoning)
13.600
Qwen3 32B (Reasoning)
13.500
GPT-5 mini (minimal)
13.400
Solar Pro 2 (Non-reasoning)
12.700
Hermes 4 - Llama-3.1 405B (Reasoning)
12.600
Qwen3 8B (Reasoning)
12.600
Qwen3 30B A3B (Reasoning)
12.100
Hermes 4 - Llama-3.1 405B (Non-reasoning)
11.800
Ling-1T
11.800
Gemini 2.5 Flash-Lite Preview (Sep '25) (Reasoning)
11.700
Hermes 4 - Llama-3.1 70B (Reasoning)
11.700
Qwen3 8B (Non-reasoning)
11.600
Sarvam 30B (high)
11.500
Ministral 3 3B
11.400
Solar Pro 2 (Reasoning)
11.400
GLM-4.5V (Reasoning)
10.900
Granite 4.1 8B
10.700
Qwen3 Omni 30B A3B (Reasoning)
10.600
Mistral Large 2 (Nov '24)
10.200
Gemini 2.5 Flash-Lite Preview (Sep '25) (Non-reasoning)
10.100
Hermes 4 - Llama-3.1 70B (Non-reasoning)
10.000
NVIDIA Nemotron 3 Nano 4B
9.800
Mistral Small 3.2
9.800
GPT-4o (Aug '24)
9.700
Qwen3 VL 32B Instruct
9.700
EXAONE 4.0 32B (Reasoning)
9.500
Qwen3 VL 30B A3B Instruct
9.500
Llama Nemotron Super 49B v1.5 (Reasoning)
9.400
NVIDIA Nemotron Nano 9B V2 (Reasoning)
9.400
Falcon-H1R-7B
9.300
GLM-4.5V (Non-reasoning)
9.200
Llama 3.3 Instruct 70B
9.100
Qwen3 4B 2507 Instruct
8.900
DeepSeek V3 (Dec '24)
8.800
Gemma 4 E4B (Non-reasoning)
8.700
Qwen3 1.7B (Reasoning)
8.700
GPT-5 nano (minimal)
8.600
NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning)
8.500
Llama Nemotron Super 49B v1.5 (Non-reasoning)
8.400
GPT-4o (Nov '24)
8.400
Mistral Small 3.1
8.400
Ling-flash-2.0
8.300
NVIDIA Nemotron Nano 9B V2 (Non-reasoning)
7.800
Qwen3 VL 4B Instruct
7.800
Llama 3.1 Nemotron Instruct 70B
7.700
Granite 4.0 1B
7.600
Gemma 4 E2B (Non-reasoning)
7.400
Qwen3 30B A3B (Non-reasoning)
7.400
Llama 4 Maverick
7.200
Nanbeige4.1-3B
7.200
Qwen3 1.7B (Non-reasoning)
7.200
NVIDIA Nemotron Nano 12B v2 VL (Reasoning)
7.100
Olmo 3.1 32B Instruct
7.100
Qwen3 30B A3B 2507 Instruct
7.100
Claude 3 Haiku
7.000
Qwen3 0.6B (Reasoning)
7.000
Gemma 4 E4B (Reasoning)
6.900
Gemma 4 E2B (Reasoning)
6.900
K2-V2 (low)
6.900
Exaone 4.0 1.2B (Non-reasoning)
6.800
LFM2.5-1.2B-Thinking
6.500
Granite 4.1 3B
6.500
Granite 4.0 H 1B
6.500
NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)
6.400
Llama 3.1 Instruct 405B
6.300
Gemini 2.5 Flash-Lite (Non-reasoning)
6.300
Gemini 2.5 Flash-Lite (Reasoning)
6.100
Granite 4.0 H Small
5.800
GPT-4.1 nano
5.800
Nova Lite
5.800
Exaone 4.0 1.2B (Reasoning)
5.500
Qwen3 Omni 30B A3B Instruct
5.500
Llama 3.1 Instruct 8B
5.500
Step3 VL 10B
5.400
Jamba Reasoning 3B
5.300
Llama 4 Scout
5.200
Command A
5.100
Llama 3.1 Instruct 70B
5.100
Llama 3.2 Instruct 11B (Vision)
4.900
Granite 4.0 H 350M
4.900
Qwen3 0.6B (Non-reasoning)
4.900
Nova Micro
4.700
Nova Pro
4.700
LFM2 2.6B
4.500
Jamba 1.7 Large
4.500
Granite 4.0 350M
4.400
Ling-mini-2.0
4.400
Apertus 70B Instruct
4.300
Olmo 3 7B Instruct
4.200
Granite 4.0 Micro
4.200
Jamba 1.7 Mini
4.200
LFM2 1.2B
4.200
Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)
3.800
Apertus 8B Instruct
3.800
DeepSeek R1 (Jan '25)
3.800
LFM2 24B A2B
3.700
LFM2.5-1.2B-Instruct
3.600
Gemma 3 12B Instruct
3.600
LFM2 8B A1B
3.500
Gemma 3 27B Instruct
3.500
Gemma 3 1B Instruct
3.500
Granite 3.3 8B (Non-reasoning)
3.500
Gemma 3 270M
3.000
LFM2.5-VL-1.6B
2.800
Phi-4 Mini Instruct
2.700
Gemma 3 4B Instruct
1.700
Gemma 3n E4B Instruct
1.700
EXAONE 4.0 32B (Non-reasoning)
1.400
Gemini 3 Deep Think
0.000
Phi-4
0.000
Olmo 3.1 32B Think
0.000
Molmo2-8B
0.000
Olmo 3 7B Think
0.000
Reka Flash 3
0.000
ERNIE 4.5 300B A47B
0.000
Tiny Aya Global
0.000
Ring-flash-2.0
0.000
Llama 3 Instruct 70B
0.000
Llama 3 Instruct 8B
0.000
Llama 3.2 Instruct 1B
0.000
Gemma 3n E2B Instruct
0.000
DeepSeek V3.2 Speciale
0.000
Olmo 3 32B Think
0.000
Sarvam M (Reasoning)
0.000

数据来源：Artificial Analysis · 定期同步