AgentUTS

Model Performance

Independent performance scores across the major model families, sourced from open leaderboards.
Source: Artificial Analysis — https://artificialanalysis.ai/leaderboards/models
264 models scored, as of 2026-10-13. Scores are the Artificial Analysis intelligence index (0–100).
Relative strength
Claude Opus 5.5 (Max, Default Fallback)Anthropic57.6/ 100
57.6
Claude Sonnet 5.5 (Max, Default Fallback)Anthropic56.0/ 100
56.0
Claude Opus 5.5 (Xhigh, Default Fallback)Anthropic56.0/ 100
56.0
Muse Wonder Full (Max)Meta55.8/ 100
55.8
Claude Opus 5.5 (High, Default Fallback)Anthropic53.6/ 100
53.6
Claude Fable 5.1 (Max, Default Fallback)Anthropic53.4/ 100
53.4
Claude Fable 5.1 (Xhigh, Default Fallback)Anthropic53.2/ 100
53.2
GPT-6 Astra (Max)OpenAI52.7/ 100
52.7
Gemini 4 Argon (High)Google52.6/ 100
52.6
GPT-6 Astra (Xhigh)OpenAI52.4/ 100
52.4
Muse Wonder (Max)Meta52.4/ 100
52.4
Claude Sonnet 5.5 (Xhigh, Default Fallback)Anthropic51.9/ 100
51.9
GPT-6.1 Sol (Max)OpenAI51.8/ 100
51.8
Claude Fable 5.1 (High, Default Fallback)Anthropic51.2/ 100
51.2
Claude Opus 5.5 (Medium, Default Fallback)Anthropic51.2/ 100
51.2
GPT-6.1 Sol (Xhigh)OpenAI51.0/ 100
51.0
GPT-6 Astra (High)OpenAI50.9/ 100
50.9
GPT-6.1 Sol (High)OpenAI50.2/ 100
50.2
GPT-6 Astra (Medium)OpenAI49.6/ 100
49.6
Claude Fable 5.1 (Medium, Default Fallback)Anthropic48.9/ 100
48.9
Muse Spark 1.3 (Max)Meta48.1/ 100
48.1
GPT-6.1 Sol (Medium)OpenAI47.8/ 100
47.8
Claude Fable 5.1 (Low, Default Fallback)Anthropic46.8/ 100
46.8
Claude Sonnet 5.5 (High, Default Fallback)Anthropic46.8/ 100
46.8
Grok 4.7 (Xhigh)SpaceXAI46.4/ 100
46.4
MiMo-V2.6-ProXiaomi46.3/ 100
46.3
Grok 4.7 (High)SpaceXAI46.3/ 100
46.3
GPT-6 Astra (Low)OpenAI45.8/ 100
45.8
Qwen3.8 Max (0902)Alibaba45.4/ 100
45.4
Muse Spark 1.3 (Xhigh)Meta45.1/ 100
45.1
GLM-5.3 (Max)Z AI44.8/ 100
44.8
Step 5 PreviewStepFun43.7/ 100
43.7
Kimi K3 (Max)Kimi43.6/ 100
43.6
Claude Haiku 5.5 (Max)Anthropic43.4/ 100
43.4
Claude Opus 5.5 (Low, Default Fallback)Anthropic42.3/ 100
42.3
Grok 4.7 (Low)SpaceXAI42.2/ 100
42.2
GPT-5.6 Terra (Max)OpenAI42.1/ 100
42.1
GPT-6.1 Sol (Low)OpenAI42.1/ 100
42.1
GLM-5.3-FlashZ AI41.8/ 100
41.8
Claude Haiku 5.5 (Xhigh)Anthropic41.2/ 100
41.2
Ling 3.1 FlashInclusionAI41.1/ 100
41.1
Gemini 3.8 Flash (High)Google40.9/ 100
40.9
Claude Sonnet 5.5 (Medium, Default Fallback)Anthropic40.8/ 100
40.8
Qwen3.8 2.4T A95BAlibaba39.9/ 100
39.9
Gemini 3.8 Flash (Medium)Google39.8/ 100
39.8
Qwen3.8-Flash-NextAlibaba39.8/ 100
39.8
DeepSeek V4.1 Flash (Max)DeepSeek39.5/ 100
39.5
Mistral Large 4 PreviewMistral38.4/ 100
38.4
GPT-6 Luna (Max)OpenAI38.1/ 100
38.1
GPT-5.6 Terra (Xhigh)OpenAI38.0/ 100
38.0
MiMo-V2.6-FlashXiaomi37.9/ 100
37.9
Claude Haiku 5.5 (High)Anthropic37.8/ 100
37.8
DeepSeek V4 Pro 0813 (Max)DeepSeek36.0/ 100
36.0
Claude Sonnet 5.5 (Low, Default Fallback)Anthropic35.9/ 100
35.9
DeepSeek V4 Flash Vision (Max)DeepSeek34.8/ 100
34.8
GPT-6 Luna (Xhigh)OpenAI34.6/ 100
34.6
Claude Haiku 5.5 (Medium)Anthropic34.5/ 100
34.5
GLM-5.3 (Low)Z AI34.3/ 100
34.3
GPT-5.6 Terra (High)OpenAI34.2/ 100
34.2
JT-4.1 Flash 236B A21B (Reasoning)China Mobile33.9/ 100
33.9
Qwen3.8 27B (Xhigh)Alibaba33.7/ 100
33.7
Motif 3Motif Technologies33.6/ 100
33.6
Gemini 3.8 Flash (Low)Google33.5/ 100
33.5
GPT-6 Luna (High)OpenAI32.9/ 100
32.9
GPT-5.3 Codex (Xhigh)OpenAI32.5/ 100
32.5
K2 Horizon 375B A23BInstitute of Foundation Models30.5/ 100
30.5
GPT-5.6 Terra (Medium)OpenAI30.1/ 100
30.1
Kimi K3 (Low)Kimi30.1/ 100
30.1
GPT-6 Luna (Medium)OpenAI29.9/ 100
29.9
Gemini 3.1 Pro PreviewGoogle29.7/ 100
29.7
Claude Haiku 5.5 (Low)Anthropic29.5/ 100
29.5
MiniMax-M3MiniMax29.2/ 100
29.2
Solar Pro 4Upstage28.2/ 100
28.2
Nex-N2-Pro (Based on Qwen3.5-397B-A17B)Nex AGI28.2/ 100
28.2
Qwen3.8 27B (Medium)Alibaba27.6/ 100
27.6
GPT-5.6 Terra (Low)OpenAI27.5/ 100
27.5
JT-4.1 Flash 236B A21B (Non-reasoning)China Mobile27.3/ 100
27.3
Quasar 438B (Max, Based on GLM-5.2)Multiverse Computing26.7/ 100
26.7
Apodex 1.1Apodex26.4/ 100
26.4
Qwen3.8 27B (Low)Alibaba26.2/ 100
26.2
GPT-5.5 Instant (June 2026)OpenAI26.0/ 100
26.0
Kimi K2.7 CodeKimi25.8/ 100
25.8
Inkling SmallThinking Machines25.7/ 100
25.7
K2 Horizon MoVA 36B A4BInstitute of Foundation Models25.3/ 100
25.3
Hy3Tencent25.3/ 100
25.3
Qwen3.7 PlusAlibaba25.2/ 100
25.2
Inkling (Xhigh)Thinking Machines25.0/ 100
25.0
DeepSeek V4.1 Flash (Non-reasoning)DeepSeek24.7/ 100
24.7
Solar Open2 250BUpstage24.7/ 100
24.7
Ling-3.0-flash-VLInclusionAI24.6/ 100
24.6
Solar Mini 4Upstage24.1/ 100
24.1
Nemotron 3 Ultra 550B A55B (Reasoning)NVIDIA22.9/ 100
22.9
Ling-3.0-flash-FinInclusionAI22.6/ 100
22.6
Gemini 3.5 Flash-LiteGoogle22.2/ 100
22.2
G9v3-39A5BAI9Stars21.8/ 100
21.8
KAT Coder Pro V2KwaiKAT21.7/ 100
21.7
A.X-K2SK Telecom21.5/ 100
21.5
GPT-6 Luna (Low)OpenAI21.5/ 100
21.5
Qwen3.5 397B A17B (Non-reasoning)Alibaba21.4/ 100
21.4
GPT-5.6 Terra (Non-reasoning)OpenAI20.8/ 100
20.8
K2 Horizon 7BInstitute of Foundation Models20.6/ 100
20.6
Qwen3.5 Omni PlusAlibaba20.4/ 100
20.4
DeepSeek V4 Pro 0813 (Non-reasoning)DeepSeek20.4/ 100
20.4
o3OpenAI20.2/ 100
20.2
Qwen3.8 27B (Non-reasoning)Alibaba20.2/ 100
20.2
K-EXAONE 2.0 0803LG AI Research19.7/ 100
19.7
LongCat 2.0LongCat19.1/ 100
19.1
JT-35B-FlashChina Mobile18.7/ 100
18.7
GPT-6 Luna (Non-reasoning)OpenAI18.5/ 100
18.5
Qwen3.5 397B A17B (Reasoning)Alibaba18.4/ 100
18.4
Qwen3.6 35B A3B (Reasoning)Alibaba18.2/ 100
18.2
Qwen3.5 122B A10B (Non-reasoning)Alibaba17.7/ 100
17.7
Muse Glimmer (High)Meta17.5/ 100
17.5
Doubao Seed CodeByteDance Seed16.9/ 100
16.9
Gemma 4 26B A4B (Reasoning)Google16.7/ 100
16.7
Ring-2.6-1TInclusionAI16.6/ 100
16.6
Qwen3.5 122B A10B (Reasoning)Alibaba15.6/ 100
15.6
K2 Horizon 3.7BInstitute of Foundation Models15.6/ 100
15.6
Qwen3.6 35B A3B (Non-reasoning)Alibaba15.2/ 100
15.2
Gemma 4 31B (Reasoning)Google14.7/ 100
14.7
ERNIE 5.0 Thinking PreviewBaidu14.3/ 100
14.3
Nova 2.0 Pro Preview (Medium)Amazon14.2/ 100
14.2
Mistral Medium 3.5Mistral14.2/ 100
14.2
Gemma 4 12B (Reasoning)Google14.2/ 100
14.2
Gemma 4 31B (Non-reasoning)Google13.9/ 100
13.9
Nova 2.0 Omni (Medium)Amazon13.6/ 100
13.6
Nova 2.0 Lite (High)Amazon13.4/ 100
13.4
Apriel-v1.6-15B-ThinkerServiceNow13.4/ 100
13.4
Qwen3.5 9B (Non-reasoning)Alibaba13.3/ 100
13.3
EXAONE 4.5 33B (Reasoning)LG AI Research13.2/ 100
13.2
Command A+Cohere13.1/ 100
13.1
Qwen3.5 4B (Reasoning)Alibaba13.1/ 100
13.1
Gemma 4 26B A4B (Non-reasoning)Google13.1/ 100
13.1
Nemotron 3.5 LightningNVIDIA12.9/ 100
12.9
Nemotron 3 Super 120B A12B (Reasoning)NVIDIA12.8/ 100
12.8
Nova 2.0 Pro Preview (Low)Amazon12.8/ 100
12.8
Granite 4.2 30BIBM12.7/ 100
12.7
MiniCPM5-2BOpenBMB12.5/ 100
12.5
Qwen3.5 Omni FlashAlibaba12.5/ 100
12.5
Nova 2.0 Lite (Medium)Amazon12.5/ 100
12.5
Mercury 2.5Inception12.3/ 100
12.3
JT-MINIChina Mobile12.2/ 100
12.2
Nova 2.0 Lite (Low)Amazon11.8/ 100
11.8
Magistral Medium 1.2Mistral11.8/ 100
11.8
HyperNova 60B 2605 (High, Based on gpt-oss-120b)Multiverse Computing11.7/ 100
11.7
Nemotron Cascade 2 30B A3BNVIDIA11.7/ 100
11.7
gpt-oss-120b (High)OpenAI11.6/ 100
11.6
K2 Think V2Institute of Foundation Models11.5/ 100
11.5
LongCat Flash LiteLongCat11.5/ 100
11.5
HyperCLOVA X SEED Think (32B)Naver11.4/ 100
11.4
Mistral Small 4 (Reasoning)Mistral11.3/ 100
11.3
Qwen3.5 9B (Reasoning)Alibaba11.2/ 100
11.2
Qwen3 Next 80B A3B (Reasoning)Alibaba11.2/ 100
11.2
Ling 3.0 TinyInclusionAI11.1/ 100
11.1
Granite 4.2 8BIBM11.1/ 100
11.1
Nova 2.0 Omni (Low)Amazon11.1/ 100
11.1
Mi:dm K 2.5 ProKorea Telecom11.0/ 100
11.0
G9v3-3BAI9Stars10.8/ 100
10.8
Qwen3.5 4B (Non-reasoning)Alibaba10.8/ 100
10.8
Trinity Large ThinkingArcee AI10.8/ 100
10.8
INTELLECT-3 (Based on GLM-4.5-Air)Prime Intellect10.6/ 100
10.6
Solar Open 100B (Reasoning)Upstage10.4/ 100
10.4
Nemotron 3 Nano Omni 30B A3B ReasoningNVIDIA10.3/ 100
10.3
gpt-oss-120b (Low)OpenAI10.2/ 100
10.2
gpt-oss-20b (Low)OpenAI10.0/ 100
10.0
Llama 4 MaverickMeta10.0/ 100
10.0
Nova 2.0 Pro Preview (Non-reasoning)Amazon10.0/ 100
10.0
North Mini CodeCohere9.9/ 100
9.9
K2-V2 (High)Institute of Foundation Models9.9/ 100
9.9
Qwen3 Next 80B A3B InstructAlibaba9.6/ 100
9.6
DiffusionGemma 26B A4BGoogle9.5/ 100
9.5
Gemma 4 12B (Non-reasoning)Google9.4/ 100
9.4
Motif-2-12.7B-ReasoningMotif Technologies9.2/ 100
9.2
Nova PremierAmazon9.2/ 100
9.2
Qwen3 Coder NextAlibaba9.2/ 100
9.2
Granite 4.2 3BIBM9.1/ 100
9.1
Mistral Small 4 (Non-reasoning)Mistral9.0/ 100
9.0
gpt-oss-20b (High)OpenAI9.0/ 100
9.0
Tri-21B-ThinkTrillion Labs9.0/ 100
9.0
K2-V2 (Medium)Institute of Foundation Models9.0/ 100
9.0
NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)NVIDIA8.9/ 100
8.9
Gemma 4 E4B (Reasoning)Google8.9/ 100
8.9
Sarvam 105B (High)Sarvam8.8/ 100
8.8
MiniCPM5-1B (Reasoning)OpenBMB8.8/ 100
8.8
Nova 2.0 Lite (Non-reasoning)Amazon8.7/ 100
8.7
MiniCPM5-1B (Non-reasoning)OpenBMB8.7/ 100
8.7
Magistral Small 1.2Mistral8.6/ 100
8.6
Nanbeige4.1-3BNanbeige8.4/ 100
8.4
LFM2.5-2.6BLiquid AI8.4/ 100
8.4
Nova 2.0 Omni (Non-reasoning)Amazon8.2/ 100
8.2
EXAONE 4.0 32B (Reasoning)LG AI Research8.2/ 100
8.2
Llama 4 ScoutMeta8.1/ 100
8.1
Hermes 4 - Llama-3.1 70B (Reasoning)Nous Research7.9/ 100
7.9
Gemma 4 E2B (Reasoning)Google7.8/ 100
7.8
Qwen3 Omni 30B A3B (Reasoning)Alibaba7.8/ 100
7.8
Falcon-H1R-7BTII UAE7.8/ 100
7.8
Llama 3.3 Instruct 70BMeta7.7/ 100
7.7
Step3 VL 10BStepFun7.7/ 100
7.7
Gemma 4 E4B (Non-reasoning)Google7.5/ 100
7.5
NVIDIA Nemotron Nano 12B v2 VL (Reasoning)NVIDIA7.5/ 100
7.5
Hermes 4 - Llama-3.1 405B (Reasoning)Nous Research7.5/ 100
7.5
ERNIE 4.5 300B A47BBaidu7.5/ 100
7.5
Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)NVIDIA7.5/ 100
7.5
Hermes 4 - Llama-3.1 405B (Non-reasoning)Nous Research7.4/ 100
7.4
NVIDIA Nemotron Nano 9B V2 (Reasoning)NVIDIA7.4/ 100
7.4
Llama 3.1 Instruct 405BMeta7.3/ 100
7.3
Kimi Linear 48B A3B InstructKimi7.3/ 100
7.3
K2-V2 (Low)Institute of Foundation Models7.3/ 100
7.3
Ring-flash-2.0InclusionAI7.2/ 100
7.2
LFM2.5-8B-A1BLiquid AI7.2/ 100
7.2
Olmo 3.1 32B ThinkAllen Institute for AI7.1/ 100
7.1
Command ACohere7.0/ 100
7.0
Qwen3.5 2B (Reasoning)Alibaba6.9/ 100
6.9
NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning)NVIDIA6.8/ 100
6.8
NVIDIA Nemotron Nano 9B V2 (Non-reasoning)NVIDIA6.8/ 100
6.8
Hermes 4 - Llama-3.1 70B (Non-reasoning)Nous Research6.7/ 100
6.7
Sarvam 30B (High)Sarvam6.6/ 100
6.6
Gemma 4 E2B (Non-reasoning)Google6.5/ 100
6.5
Olmo 3.1 32B InstructAllen Institute for AI6.5/ 100
6.5
R1 1776Perplexity6.4/ 100
6.4
Llama 3.2 Instruct 90B (Vision)Meta6.4/ 100
6.4
EXAONE 4.0 32B (Non-reasoning)LG AI Research6.3/ 100
6.3
Celeris-1Celeris6.3/ 100
6.3
Phi-4 Mini InstructMicrosoft6.3/ 100
6.3
Qwen3.5 2B (Non-reasoning)Alibaba6.2/ 100
6.2
Jamba 1.7 LargeAI21 Labs6.1/ 100
6.1
Qwen3.5 0.8B (Reasoning)Alibaba6.1/ 100
6.1
DeepHermes 3 - Mistral 24B Preview (Non-reasoning)Nous Research6.1/ 100
6.1
Qwen3 Omni 30B A3B InstructAlibaba6.0/ 100
6.0
Granite 4.0 H SmallIBM6.0/ 100
6.0
Ministral 3 14BMistral6.0/ 100
6.0
Nova MicroAmazon5.9/ 100
5.9
Phi-4Microsoft5.9/ 100
5.9
LFM2 24B A2BLiquid AI5.9/ 100
5.9
Phi-4 Multimodal InstructMicrosoft5.8/ 100
5.8
NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)NVIDIA5.8/ 100
5.8
MiniCPM-V 4.6 1.3BOpenBMB5.7/ 100
5.7
Jamba Reasoning 3BAI21 Labs5.7/ 100
5.7
Olmo 3 7B ThinkAllen Institute for AI5.6/ 100
5.6
Molmo 7B-DAllen Institute for AI5.6/ 100
5.6
Reka Flash 3Reka AI5.6/ 100
5.6
Ministral 3 8BMistral5.5/ 100
5.5
Qwen3.5 0.8B (Non-reasoning)Alibaba5.4/ 100
5.4
Llama 3.2 Instruct 11B (Vision)Meta5.4/ 100
5.4
Exaone 4.0 1.2B (Reasoning)LG AI Research5.3/ 100
5.3
LFM2.5-1.2B-InstructLiquid AI5.2/ 100
5.2
LFM2.5-1.2B-ThinkingLiquid AI5.2/ 100
5.2
Exaone 4.0 1.2B (Non-reasoning)LG AI Research5.2/ 100
5.2
Olmo 3 7B InstructAllen Institute for AI5.2/ 100
5.2
Jamba 1.7 MiniAI21 Labs5.2/ 100
5.2
Granite 4.0 H 1BIBM5.2/ 100
5.2
Gemma 3 270MGoogle5.1/ 100
5.1
Granite 4.0 MicroIBM5.1/ 100
5.1
Apertus 70B InstructSwiss AI Initiative5.1/ 100
5.1
DeepHermes 3 - Llama-3.1 8B Preview (Non-reasoning)Nous Research5.1/ 100
5.1
Molmo2-8BAllen Institute for AI5.0/ 100
5.0
LFM2.5-VL-1.6BLiquid AI4.8/ 100
4.8
Granite 4.0 350MIBM4.8/ 100
4.8
Ministral 3 3BMistral4.8/ 100
4.8
Tiny Aya GlobalCohere4.8/ 100
4.8
Apertus 8B InstructSwiss AI Initiative4.8/ 100
4.8
NVIDIA Nemotron 3 Nano 4BNVIDIA4.8/ 100
4.8
Granite 4.0 H 350MIBM4.8/ 100
4.8
K2 Horizon 0.9BInstitute of Foundation Models3.0/ 100
3.0
264 models Scores as of 2026-10 Compare with pricing →