Benchmarks / Artificial Analysis
Reported by Artificial Analysis
Artificial Analysis
Share of expert-written exam questions answered correctly, as run by Artificial Analysis with its own harness and prompts.
- Results dated
- 1 Oct 2026
- Models
- 314
- Unit
- % of questions
Full results
| # | Model | Artificial Analysis: Humanity's Last Exam % of questions, higher is better |
|---|---|---|
| 1 | Claude Opus 5.5 (max reasoning)Anthropic |
61.4%
|
| 2 | Claude Fable 5.1 (max)Anthropic |
59.1%
|
| 3 | Claude Fable 5.1 (xhigh)Anthropic |
58.7%
|
| 4 | Claude Opus 5.5 (xhigh reasoning)Anthropic |
57.5%
|
| 5 | Gemini 4 Argon (high)Google |
57.1%
|
| 6 | Claude Fable 5.1 (high reasoning)Anthropic |
55.9%
|
| 7 | Claude Opus 5.5 (high)Anthropic |
55.6%
|
| 8 | Claude Fable 5 (max reasoning)Anthropic |
55.5%
|
| 9 | Claude Sonnet 5.5 (max)Anthropic |
55.0%
|
| 10 | Claude Opus 5 (max)Anthropic |
54.9%
|
| 11 | Claude Opus 5.5 (medium reasoning)Anthropic |
54.7%
|
| 11 | GPT-6 Astra (max)OpenAI |
54.7%
|
| 13 | GPT-6 Astra (xhigh reasoning)OpenAI |
54.6%
|
| 14 | Claude Opus 5 (xhigh)Anthropic |
54.4%
|
| 15 | Claude Fable 5.1 (medium reasoning)Anthropic |
53.8%
|
| 16 | GPT-6 Astra (high reasoning)OpenAI |
53.1%
|
| 17 | GPT-6.1 Sol (max)OpenAI |
52.9%
|
| 18 | Claude Opus 5 (high)Anthropic |
52.8%
|
| 19 | GPT-6 Astra (medium reasoning)OpenAI |
52.7%
|
| 20 | GPT-6.1 Sol (xhigh)OpenAI |
52.6%
|
| 21 | GPT-6.1 Sol (high)OpenAI |
51.4%
|
| 22 | Claude Opus 5 (medium)Anthropic |
51.3%
|
| 23 | Claude Sonnet 5.5 (xhigh)Anthropic |
50.0%
|
| 24 | GPT-6.1 Sol (medium)OpenAI |
49.9%
|
| 25 | GPT-5.6 Sol (max)OpenAI |
49.5%
|
| 26 | GPT-6 Astra (low reasoning)OpenAI |
49.2%
|
| 27 | Claude Fable 5.1 (low reasoning)Anthropic |
48.9%
|
| 28 | Claude Opus 4.8 (max reasoning)Anthropic |
48.7%
|
| 28 | Muse Spark 1.3 (max)Meta |
48.7%
|
| 30 | Claude Opus 5.5 (low reasoning)Anthropic |
48.3%
|
| 31 | Gemini 3.7 Flash (high)Google |
47.9%
|
| 31 | GPT-6 Sol (max)OpenAI |
47.9%
|
| 33 | Gemini 3.8 Flash (high)Google |
47.8%
|
| 34 | Muse Spark 1.3 (xhigh)Meta |
47.5%
|
| 35 | GPT-6.1 Sol (low)OpenAI |
47.4%
|
| 36 | GPT-5.6 Sol (xhigh)OpenAI |
47.3%
|
| 37 | Gemini 3.1 Pro PreviewGoogle |
47.0%
|
| 38 | Kimi K3 (max)Moonshot AI |
46.9%
|
| 39 | GPT-6 Sol (xhigh)OpenAI |
46.3%
|
| 40 | Muse Spark 1.1 (xhigh)Meta |
46.2%
|
| 41 | GPT-5.6 Sol (high reasoning)OpenAI |
46.0%
|
| 42 | Claude Sonnet 5.5 (high)Anthropic |
45.8%
|
| 42 | GPT-5.5 (xhigh reasoning)OpenAI |
45.8%
|
| 44 | Muse Spark 1.2Meta |
45.5%
|
| 45 | GPT-5.5 (high)OpenAI |
45.0%
|
| 46 | GPT-6 Sol (high)OpenAI |
44.1%
|
| 46 | Grok 4.6 (xhigh)xAI |
44.1%
|
| 48 | GPT-5.4 (xhigh reasoning)OpenAI |
43.7%
|
| 49 | Claude Opus 5 (low)Anthropic |
43.4%
|
| 50 | Qwen3.8 Max (0902)Alibaba |
43.1%
|
| 50 | Grok 4.7 (xhigh)xAI |
43.1%
|
| 52 | Qwen3.8 Max (0803)Alibaba |
43.0%
|
| 53 | GPT-5.6 Terra (max)OpenAI |
42.9%
|
| 53 | Grok 4.6 (high)xAI |
42.9%
|
| 55 | Gemini 3.5 Flash (high)Google |
42.7%
|
| 55 | Grok 4.5 (high)xAI |
42.7%
|
| 57 | GPT-5.3-Codex (xhigh)OpenAI |
42.5%
|
| 58 | Qwen3.8 2.4T A95BAlibaba |
42.4%
|
| 58 | GPT-5.5 (medium reasoning)OpenAI |
42.4%
|
| 60 | Claude Opus 4.7 (max reasoning)Anthropic |
42.3%
|
| 60 | Grok 4.7 (high)xAI |
42.3%
|
| 60 | GLM 5.3 (max)Z.ai |
42.3%
|
| 63 | GPT-5.6 Sol (medium reasoning)OpenAI |
42.2%
|
| 64 | Gemini 3.8 Flash (medium reasoning)Google |
42.1%
|
| 64 | Grok 4.6 (medium)xAI |
42.1%
|
| 66 | GPT-5.6 Terra (xhigh)OpenAI |
41.9%
|
| 67 | Claude Sonnet 5 (max reasoning)Anthropic |
41.3%
|
| 67 | Gemini 3.5 Flash (medium)Google |
41.3%
|
| 69 | GLM 5.2 (max)Z.ai |
41.1%
|
| 70 | DeepSeek V4 Pro 0423 (max)DeepSeek |
41.0%
|
| 70 | GPT-6 Sol (medium)OpenAI |
41.0%
|
| 72 | Gemini 3.6 Flash (high)Google |
40.8%
|
| 73 | muse-sparkMeta |
40.7%
|
| 74 | Qwen3.7 MaxAlibaba |
40.5%
|
| 75 | Claude Opus 4.6 (max reasoning)Anthropic |
39.9%
|
| 75 | GLM 5.3 FlashZ.ai |
39.9%
|
| 77 | Claude Sonnet 5.5 (medium)Anthropic |
39.8%
|
| 78 | gemini-3-pro (high)Google |
39.7%
|
| 79 | GPT-5.6 Luna (max)OpenAI |
39.5%
|
| 80 | GPT-5.6 Sol (low reasoning)OpenAI |
39.4%
|
| 81 | DeepSeek V4.1 Flash (max)DeepSeek |
39.2%
|
| 82 | Claude Sonnet 5 (xhigh reasoning)Anthropic |
39.0%
|
| 82 | Gemini 3.7 Flash (medium reasoning)Google |
39.0%
|
| 84 | DeepSeek V4 Flash 0423 (max)DeepSeek |
38.6%
|
| 85 | GPT-5.6 Terra (high reasoning)OpenAI |
38.5%
|
| 85 | GPT-6 Luna (max)OpenAI |
38.5%
|
| 87 | Grok Build 0.1xAI |
38.3%
|
| 88 | Qwen3.8 Flash NextAlibaba |
38.0%
|
| 89 | GPT-5.2 (xhigh)OpenAI |
37.7%
|
| 90 | deepseek-v4-pro-preview (max reasoning)DeepSeek |
37.5%
|
| 90 | Kimi K2.6Moonshot AI |
37.5%
|
| 92 | Grok 4.3 (high)xAI |
37.2%
|
| 93 | Gemini 3.8 Flash (low reasoning)Google |
37.1%
|
| 94 | GPT-5.6 Luna (xhigh)OpenAI |
37.0%
|
| 95 | gemini-3-flash (thinking reasoning)Google |
36.6%
|
| 95 | GLM 5.3 (low reasoning)Z.ai |
36.6%
|
| 97 | Claude Sonnet 5.5 (low reasoning)Anthropic |
36.2%
|
| 98 | Claude Sonnet 5 (high)Anthropic |
35.7%
|
| 98 | GPT-5.2-Codex (xhigh)OpenAI |
35.7%
|
| 100 | Qwen3.7 PlusAlibaba |
35.6%
|
| 101 | deepseek-v4-pro-preview (high reasoning)DeepSeek |
35.2%
|
| 102 | Gemini 3.7 Flash (low reasoning)Google |
35.1%
|
| 103 | Kimi K2.7 CodeMoonshot AI |
35.0%
|
| 104 | GPT-6 Sol (low)OpenAI |
34.9%
|
| 105 | deepseek-v4-flash-preview (max reasoning)DeepSeek |
34.8%
|
| 106 | DeepSeek V4 Flash Vision (experimental, max)DeepSeek |
34.5%
|
| 106 | Grok 4.20xAI |
34.5%
|
| 108 | GPT-6 Luna (xhigh)OpenAI |
34.3%
|
| 109 | Qwen3.8 27B (xhigh)Alibaba |
33.9%
|
| 110 | Claude Sonnet 4.6 (max reasoning)Anthropic |
33.6%
|
| 111 | GPT-5.6 Luna (high reasoning)OpenAI |
33.4%
|
| 112 | Claude Opus 4.7 (off)Anthropic |
33.3%
|
| 112 | GPT-5.6 Terra (medium reasoning)OpenAI |
33.3%
|
| 114 | GPT-6 Luna (high)OpenAI |
32.9%
|
| 115 | GPT-5.5 (low reasoning)OpenAI |
32.7%
|
| 116 | grok-4.20-0309-reasoningxAI |
32.4%
|
| 117 | Qwen3.6 Max PreviewAlibaba |
30.8%
|
| 117 | GPT-5.4 (low reasoning)OpenAI |
30.8%
|
| 119 | Kimi K2.5Moonshot AI |
30.7%
|
| 120 | deepseek-v4-flash-preview (high reasoning)DeepSeek |
30.3%
|
| 121 | Claude Opus 4.5 (thinking reasoning)Anthropic |
30.1%
|
| 121 | GLM 5.1Z.ai |
30.1%
|
| 123 | Claude Sonnet 5 (medium reasoning)Anthropic |
30.0%
|
| 123 | Grok 4.3 (medium)xAI |
30.0%
|
| 125 | gemini-3-pro (low)Google |
29.5%
|
| 126 | GLM 5Z.ai |
29.3%
|
| 127 | GPT-5.6 Terra (low reasoning)OpenAI |
29.2%
|
| 128 | Qwen3.5 397B A17BAlibaba |
29.0%
|
| 129 | deepseek-v3.2-specialeDeepSeek |
28.7%
|
| 130 | GPT-5.1 (high)OpenAI |
28.5%
|
| 130 | GPT-5 (high)OpenAI |
28.5%
|
| 132 | GPT-5.4 Nano (xhigh)OpenAI |
28.3%
|
| 132 | GPT-6 Luna (medium)OpenAI |
28.3%
|
| 134 | GPT-5.4 Mini (xhigh)OpenAI |
28.1%
|
| 135 | Qwen3 Max ThinkingAlibaba |
28.0%
|
| 136 | Grok 4.20 (off)xAI |
27.9%
|
| 136 | GLM 5.1 (no reasoning)Z.ai |
27.9%
|
| 138 | Qwen3.6 PlusAlibaba |
27.8%
|
| 138 | GPT-5-Codex (high)OpenAI |
27.8%
|
| 138 | GLM 5 TurboZ.ai |
27.8%
|
| 141 | Grok 4.6 (low)xAI |
27.6%
|
| 142 | GLM 4.7Z.ai |
27.4%
|
| 143 | GPT-5.2 (medium reasoning)OpenAI |
26.7%
|
| 143 | grok-4xAI |
26.7%
|
| 145 | GPT-5.6 Luna (medium reasoning)OpenAI |
25.8%
|
| 146 | GPT-5.1-Codex (high)OpenAI |
25.7%
|
| 147 | GPT-5 (medium)OpenAI |
25.4%
|
| 148 | Qwen3.5-122B-A10BAlibaba |
25.2%
|
| 149 | Kimi K3 (low reasoning)Moonshot AI |
25.0%
|
| 150 | DeepSeek V3.2DeepSeek |
24.6%
|
| 151 | grok-4.20-0309-non-reasoning (no reasoning)xAI |
24.5%
|
| 152 | Gemini 3.5 Flash (minimal reasoning)Google |
24.1%
|
| 153 | Qwen3.5-27BAlibaba |
23.9%
|
| 154 | Kimi K2 ThinkingMoonshot AI |
23.8%
|
| 155 | Gemma 4 31BGoogle |
23.6%
|
| 156 | Qwen3.6 27BAlibaba |
23.1%
|
| 157 | Qwen3.6 35B A3BAlibaba |
22.2%
|
| 158 | muse-glimmer (high)Meta |
22.0%
|
| 159 | Claude Sonnet 5 (low reasoning)Anthropic |
21.9%
|
| 160 | GPT-5.5 Instant (26 May)OpenAI |
21.6%
|
| 161 | GPT-5 Mini (high)OpenAI |
21.5%
|
| 162 | Qwen3.5-35B-A3BAlibaba |
21.0%
|
| 163 | GPT-6 Luna (low)OpenAI |
20.3%
|
| 164 | GPT-5.5 Instant (26 June)OpenAI |
19.9%
|
| 165 | Qwen3.5 397B A17B (no reasoning)Alibaba |
19.8%
|
| 165 | GPT-5.6 Luna (low reasoning)OpenAI |
19.8%
|
| 167 | Kimi K2.6 (no reasoning)Moonshot AI |
19.6%
|
| 167 | GPT-5 (low reasoning)OpenAI |
19.6%
|
| 167 | gpt-oss-120b (high)OpenAI |
19.6%
|
| 170 | Gemma 4 26B A4BGoogle |
19.3%
|
| 170 | grok-4-1-fast-reasoningxAI |
19.3%
|
| 172 | Claude Opus 4.6 (off)Anthropic |
19.1%
|
| 172 | grok-4-fast-reasoningxAI |
19.1%
|
| 174 | Claude Sonnet 5 (off)Anthropic |
19.0%
|
| 175 | Gemini 3.5 Flash LiteGoogle |
18.8%
|
| 176 | GPT-5.4 Mini (medium)OpenAI |
18.6%
|
| 177 | GPT-5.1-Codex-Mini (high)OpenAI |
18.5%
|
| 178 | GPT-6 Sol (no reasoning)OpenAI |
18.4%
|
| 178 | Grok 4.3 (low)xAI |
18.4%
|
| 180 | Claude Sonnet 4.5 (thinking reasoning)Anthropic |
17.8%
|
| 181 | Gemini 3.1 Flash Lite PreviewGoogle |
17.2%
|
| 182 | GLM 5V TurboZ.ai |
17.1%
|
| 183 | GPT-5.6 Sol (no reasoning)OpenAI |
16.7%
|
| 184 | DeepSeek V3.1 TerminusDeepSeek |
16.4%
|
| 185 | Qwen3 235B A22B Thinking 2507Alibaba |
15.9%
|
| 185 | Qwen3.5-122B-A10B (no reasoning)Alibaba |
15.9%
|
| 185 | GPT-5.4 Nano (medium)OpenAI |
15.9%
|
| 185 | GPT-5 Mini (medium)OpenAI |
15.9%
|
| 189 | gemma-4-12b-itGoogle |
15.7%
|
| 190 | Qwen3.6 27B (no reasoning)Alibaba |
15.1%
|
| 191 | gemini-3-flash (off)Google |
15.0%
|
| 192 | Qwen3.5-9BAlibaba |
14.9%
|
| 192 | Qwen3.5 Omni PlusAlibaba |
14.9%
|
| 192 | DeepSeek V3.2 ExpDeepSeek |
14.9%
|
| 195 | GLM 4.6 (thinking reasoning)Z.ai |
14.5%
|
| 196 | DeepSeek V3.1DeepSeek |
14.3%
|
| 197 | Qwen3.8 27B (medium)Alibaba |
14.1%
|
| 198 | Qwen3.8 27B (low)Alibaba |
14.0%
|
| 199 | Qwen3.5-27B (no reasoning)Alibaba |
13.9%
|
| 199 | Qwen3.6 35B A3B (no reasoning)Alibaba |
13.9%
|
| 201 | gemini-2.5-flash-preview-09-2025 (thinking reasoning)Google |
13.8%
|
| 201 | Mistral Medium 3.5Mistral |
13.8%
|
| 203 | GPT-5.5 (no reasoning)OpenAI |
13.7%
|
| 204 | Qwen3.5-35B-A3B (no reasoning)Alibaba |
13.4%
|
| 205 | Claude Sonnet 4.6 (off)Anthropic |
13.3%
|
| 206 | Claude Opus 4.5 (no reasoning)Anthropic |
13.2%
|
| 206 | Kimi K2.5Moonshot AI |
13.2%
|
| 208 | GLM 4.5Z.ai |
13.0%
|
| 209 | Qwen3 Max Thinking (preview)Alibaba |
12.7%
|
| 210 | Qwen3 Next 80B A3B ThinkingAlibaba |
12.6%
|
| 211 | Claude Opus 4.1 (thinking reasoning)Anthropic |
12.5%
|
| 212 | Qwen3.8 27B (off)Alibaba |
12.1%
|
| 213 | Qwen3 MaxAlibaba |
11.9%
|
| 213 | Qwen3 VL 235B A22B ThinkingAlibaba |
11.9%
|
| 215 | Gemma 4 31B (off)Google |
11.8%
|
| 216 | Gemma 4 26B A4B (off)Google |
11.5%
|
| 217 | GPT-5.6 Terra (no reasoning)OpenAI |
11.4%
|
| 218 | GPT-5.4 (no reasoning)OpenAI |
11.3%
|
| 219 | DeepSeek V3.2 (no reasoning)DeepSeek |
11.2%
|
| 220 | Qwen3 235B A22B Instruct 2507Alibaba |
11.1%
|
| 221 | gpt-oss-20b (high reasoning)OpenAI |
11.0%
|
| 222 | DeepSeek V4.1 Flash (off)DeepSeek |
10.8%
|
| 222 | DiffusionGemma 26B A4BGoogle |
10.8%
|
| 224 | DeepSeek V4 Pro 0423 (no reasoning)DeepSeek |
10.6%
|
| 225 | Claude Haiku 4.5 (thinking reasoning)Anthropic |
10.4%
|
| 226 | Qwen3 30B A3B Thinking 2507Alibaba |
10.3%
|
| 226 | Magistral Medium 1.2Mistral |
10.3%
|
| 228 | Qwen3 Coder NextAlibaba |
10.1%
|
| 228 | qwen3-max-previewAlibaba |
10.1%
|
| 228 | qwen3-vl-32b-thinkingAlibaba |
10.1%
|
| 231 | qwen3.5-4b (thinking reasoning)Alibaba |
9.9%
|
| 231 | Mistral Small 4Mistral |
9.9%
|
| 233 | GLM 5.2 (no reasoning)Z.ai |
9.8%
|
| 234 | GLM 4.6V (thinking reasoning)Z.ai |
9.6%
|
| 235 | GPT-5 Nano (high)OpenAI |
9.5%
|
| 236 | Qwen3.5-9B (no reasoning)Alibaba |
9.4%
|
| 237 | DeepSeek V3.2 Exp (no reasoning)DeepSeek |
9.0%
|
| 238 | Qwen3 VL 30B A3B ThinkingAlibaba |
8.9%
|
| 239 | DeepSeek V3.1 Terminus (no reasoning)DeepSeek |
8.7%
|
| 239 | gemini-2.5-flash-preview-09-2025 (no reasoning)Google |
8.7%
|
| 239 | GPT-5 Nano (medium)OpenAI |
8.7%
|
| 242 | GPT-6 Luna (no reasoning)OpenAI |
8.6%
|
| 243 | deepseek-v4-pro-preview (no reasoning)DeepSeek |
8.2%
|
| 244 | qwen3.5-4b (no reasoning)Alibaba |
8.0%
|
| 244 | GPT-5.2 (no reasoning)OpenAI |
8.0%
|
| 244 | Grok Code Fast 1xAI |
8.0%
|
| 247 | deepseek-v4-flash-preview (no reasoning)DeepSeek |
7.8%
|
| 248 | Qwen3.5 Omni FlashAlibaba |
7.6%
|
| 248 | Qwen3 Next 80B A3B InstructAlibaba |
7.6%
|
| 248 | GLM 4.7 FlashZ.ai |
7.6%
|
| 248 | GLM 5 (no reasoning)Z.ai |
7.6%
|
| 252 | qwen3-omni-30b-a3b-thinkingAlibaba |
7.5%
|
| 253 | Kimi K2 0711Moonshot AI |
7.4%
|
| 254 | Claude Sonnet 4.5 (no reasoning)Anthropic |
7.2%
|
| 254 | GPT-5.6 Luna (no reasoning)OpenAI |
7.2%
|
| 256 | gemini-2.5-flash-lite-preview-09-2025 (thinking reasoning)Google |
7.0%
|
| 256 | GLM 4.5 AirZ.ai |
7.0%
|
| 258 | Qwen3 30B A3B Instruct 2507Alibaba |
6.9%
|
| 259 | Qwen3 VL 32B InstructAlibaba |
6.8%
|
| 259 | Grok 4.3 (off)xAI |
6.8%
|
| 261 | DeepSeek V3.1 (no reasoning)DeepSeek |
6.7%
|
| 262 | Qwen3 VL 235B A22B InstructAlibaba |
6.6%
|
| 262 | gpt-5-chatOpenAI |
6.6%
|
| 264 | magistral-small-2509Mistral |
6.4%
|
| 264 | Kimi K2 0905Moonshot AI |
6.4%
|
| 264 | GLM 4.7 (no reasoning)Z.ai |
6.4%
|
| 267 | Qwen3 VL 30B A3B InstructAlibaba |
6.3%
|
| 267 | gemma-4-12b-it (no reasoning)Google |
6.3%
|
| 267 | GLM 4.5V (thinking reasoning)Z.ai |
6.3%
|
| 270 | qwen3-4b-thinking-2507Alibaba |
6.2%
|
| 271 | GPT-5 (minimal reasoning)OpenAI |
6.0%
|
| 272 | GPT-5.4 Mini (no reasoning)OpenAI |
5.9%
|
| 272 | gpt-oss-120b (low)OpenAI |
5.9%
|
| 274 | GLM 4.6 (no reasoning)Z.ai |
5.5%
|
| 275 | Ministral 3 3B 2512Mistral |
5.4%
|
| 276 | GPT-5.1 (off)OpenAI |
5.3%
|
| 276 | gpt-oss-20b (low reasoning)OpenAI |
5.3%
|
| 278 | qwen3.5-0.8b (no reasoning)Alibaba |
5.1%
|
| 278 | gemini-2.5-flash-lite-preview-09-2025 (no reasoning)Google |
5.1%
|
| 278 | GPT-5 Mini (minimal)OpenAI |
5.1%
|
| 278 | grok-4-1-fast-non-reasoning (no reasoning)xAI |
5.1%
|
| 282 | qwen3.5-2b (no reasoning)Alibaba |
5.0%
|
| 282 | GLM 4.7 Flash (no reasoning)Z.ai |
5.0%
|
| 284 | gemma-4-e2b-itGoogle |
4.8%
|
| 284 | gemma-4-e4b-it (no reasoning)Google |
4.8%
|
| 286 | gemma-4-e2b-it (no reasoning)Google |
4.7%
|
| 286 | Mistral Medium 3.1Mistral |
4.7%
|
| 288 | Qwen3 Omni 30B A3B InstructAlibaba |
4.6%
|
| 288 | qwen3-vl-4b-thinkingAlibaba |
4.6%
|
| 288 | Ministral 3 14B 2512Mistral |
4.6%
|
| 291 | qwen3-4b-instruct-2507Alibaba |
4.5%
|
| 291 | Qwen3 Coder 480B A35BAlibaba |
4.5%
|
| 291 | grok-4-fast-non-reasoning (no reasoning)xAI |
4.5%
|
| 294 | Ministral 3 8B 2512Mistral |
4.3%
|
| 295 | Claude Haiku 4.5 (no reasoning)Anthropic |
4.2%
|
| 295 | Mistral Large 3 2512Mistral |
4.2%
|
| 297 | GPT-5.4 Nano (no reasoning)OpenAI |
4.1%
|
| 298 | GPT-5 Nano (minimal)OpenAI |
4.0%
|
| 299 | Qwen3 Coder 30B A3B InstructAlibaba |
3.8%
|
| 299 | Qwen3 VL 8B ThinkingAlibaba |
3.8%
|
| 299 | gemma-4-e4b-itGoogle |
3.8%
|
| 299 | Devstral MediumMistral |
3.8%
|
| 299 | Devstral Small 1.1Mistral |
3.8%
|
| 299 | Mistral Small 4 (no reasoning)Mistral |
3.8%
|
| 305 | GLM 4.6V (no reasoning)Z.ai |
3.7%
|
| 306 | qwen3-vl-4b-instructAlibaba |
3.6%
|
| 306 | Gemma 3 270MGoogle |
3.6%
|
| 306 | Devstral 2 2512Mistral |
3.6%
|
| 309 | devstral-small-2-24b-instruct-2512Mistral |
3.5%
|
| 309 | GLM 4.5V (off)Z.ai |
3.5%
|
| 311 | Qwen3 VL 8B InstructAlibaba |
2.7%
|
| 312 | qwen3.5-2b (thinking reasoning)Alibaba |
2.6%
|
| 313 | kimi-linear-48b-a3b-instructMoonshot AI |
2.5%
|
| 314 | qwen3.5-0.8b (thinking reasoning)Alibaba |
1.1%
|
Results as published by Artificial Analysis; we do not re-run them.
What it measures
Share of expert-written exam questions answered correctly, as run by Artificial Analysis with its own harness and prompts.
What it does not measure
Not everyday work; academic questions at the edge of expertise.
Source
Failures
Failures count against a model: a product with no usable output is a failed listing. They are listed here so you can see why.
Not ranked
- Command A: creator outside the providers the catalogue covers
- GPT-4 Turbo: released before 2025-07-01
- GPT-4o (2024-05-13): released before 2025-07-01
- GPT-4o (2024-08-06): released before 2025-07-01
- Hy3: creator outside the providers the catalogue covers
- Hy3 preview: creator outside the providers the catalogue covers
- Inkling: creator outside the providers the catalogue covers
- Inkling Small: creator outside the providers the catalogue covers
- Llama 4 Maverick: released before 2025-07-01
- Llama 4 Scout: released before 2025-07-01
- Mercury 2: creator outside the providers the catalogue covers
- MiniMax M2: creator outside the providers the catalogue covers
- MiniMax M3: creator outside the providers the catalogue covers
- Mistral Large 2407: released before 2025-07-01
- Mistral Medium 3: released before 2025-07-01
- Phi 4: creator outside the providers the catalogue covers
- QwQ-32B-Preview: released before 2025-07-01
- R1: released before 2025-07-01
- Reka Flash 3: creator outside the providers the catalogue covers
- Solar Pro 3: creator outside the providers the catalogue covers
- Solar Pro 4: creator outside the providers the catalogue covers
- Trinity Large Thinking: creator outside the providers the catalogue covers
- a-x-k2: creator outside the providers the catalogue covers
- apertus-70b-instruct: creator outside the providers the catalogue covers
- apertus-8b-instruct: creator outside the providers the catalogue covers
- apodex-1-1: creator outside the providers the catalogue covers
- apriel-v1-5-15b-thinker: creator outside the providers the catalogue covers
- apriel-v1-6-15b-thinker: creator outside the providers the catalogue covers
- arctic-instruct: creator outside the providers the catalogue covers
- celeris-1: creator outside the providers the catalogue covers
- claude-2: released before 2025-07-01
- claude-21: released before 2025-07-01
- claude-3-5-haiku: released before 2025-07-01
- claude-3-7-sonnet: released before 2025-07-01
- claude-3-7-sonnet-thinking: released before 2025-07-01
- claude-3-haiku: released before 2025-07-01
- claude-3-opus: released before 2025-07-01
- claude-3-sonnet: released before 2025-07-01
- claude-35-sonnet: released before 2025-07-01
- claude-35-sonnet-june-24: released before 2025-07-01
- claude-4-opus: released before 2025-07-01
- claude-4-opus-thinking: released before 2025-07-01
- claude-4-sonnet: released before 2025-07-01
- claude-4-sonnet-thinking: released before 2025-07-01
- claude-instant: released before 2025-07-01
- claude-sonnet-4-6-non-reasoning-low-effort: non-standard setting combination
- cogito-v2-1-reasoning: creator outside the providers the catalogue covers
- command-a-plus: creator outside the providers the catalogue covers
- command-r-03-2024: creator outside the providers the catalogue covers
- command-r-plus-04-2024: creator outside the providers the catalogue covers
- dbrx: creator outside the providers the catalogue covers
- deephermes-3-llama-3-1-8b-preview: creator outside the providers the catalogue covers
- deephermes-3-mistral-24b-preview: creator outside the providers the catalogue covers
- deepseek-coder-v2: released before 2025-07-01
- deepseek-coder-v2-lite: released before 2025-07-01
- deepseek-llm-67b-chat: released before 2025-07-01
- deepseek-r1-0120: released before 2025-07-01
- deepseek-r1-distill-llama-70b: released before 2025-07-01
- deepseek-r1-distill-llama-8b: released before 2025-07-01
- deepseek-r1-distill-qwen-1-5b: released before 2025-07-01
- deepseek-r1-distill-qwen-14b: released before 2025-07-01
- deepseek-r1-distill-qwen-32b: released before 2025-07-01
- deepseek-r1-qwen3-8b: released before 2025-07-01
- deepseek-v2: released before 2025-07-01
- deepseek-v2-5: released before 2025-07-01
- deepseek-v2-5-sep-2024: released before 2025-07-01
- deepseek-v3: released before 2025-07-01
- deepseek-v3-0324: released before 2025-07-01
- devstral-small-2505: released before 2025-07-01
- doubao-seed-code: creator outside the providers the catalogue covers
- ernie-4-5-300b-a47b: creator outside the providers the catalogue covers
- ernie-5-0-thinking-preview: creator outside the providers the catalogue covers
- exaone-4-0-1-2b: creator outside the providers the catalogue covers
- exaone-4-0-1-2b-reasoning: creator outside the providers the catalogue covers
- exaone-4-0-32b: creator outside the providers the catalogue covers
- exaone-4-0-32b-reasoning: creator outside the providers the catalogue covers
- exaone-4-5-33b: creator outside the providers the catalogue covers
- falcon-h1r-7b: creator outside the providers the catalogue covers
- g9v3-39a5b: creator outside the providers the catalogue covers
- g9v3-3b: creator outside the providers the catalogue covers
- gemini-1-0-pro: released before 2025-07-01
- gemini-1-0-ultra: released before 2025-07-01
- gemini-1-5-flash: released before 2025-07-01
- gemini-1-5-flash-8b: released before 2025-07-01
- gemini-1-5-flash-may-2024: released before 2025-07-01
- gemini-1-5-pro: released before 2025-07-01
- gemini-1-5-pro-may-2024: released before 2025-07-01
- gemini-2-0-flash: released before 2025-07-01
- gemini-2-0-flash-experimental: released before 2025-07-01
- gemini-2-0-flash-lite-001: released before 2025-07-01
- gemini-2-0-flash-lite-preview: released before 2025-07-01
- gemini-2-0-flash-thinking-exp-0121: released before 2025-07-01
- gemini-2-0-flash-thinking-exp-1219: released before 2025-07-01
- gemini-2-0-pro-experimental-02-05: released before 2025-07-01
- gemini-2-5-flash: released before 2025-07-01
- gemini-2-5-flash-04-2025: released before 2025-07-01
- gemini-2-5-flash-lite: released before 2025-07-01
- gemini-2-5-flash-lite-reasoning: released before 2025-07-01
- gemini-2-5-flash-reasoning: released before 2025-07-01
- gemini-2-5-flash-reasoning-04-2025: released before 2025-07-01
- gemini-2-5-pro: released before 2025-07-01
- gemini-2-5-pro-03-25: released before 2025-07-01
- gemini-2-5-pro-05-06: released before 2025-07-01
- gemma-3-12b: released before 2025-07-01
- gemma-3-1b: released before 2025-07-01
- gemma-3-27b: released before 2025-07-01
- gemma-3-4b: released before 2025-07-01
- gemma-3n-e2b: released before 2025-07-01
- gemma-3n-e4b: released before 2025-07-01
- gemma-3n-e4b-preview-0520: released before 2025-07-01
- gpt-35-turbo: released before 2025-07-01
- gpt-4: released before 2025-07-01
- gpt-4-1: released before 2025-07-01
- gpt-4-1-mini: released before 2025-07-01
- gpt-4-1-nano: released before 2025-07-01
- gpt-4-5: released before 2025-07-01
- gpt-4o: released before 2025-07-01
- gpt-4o-chatgpt: released before 2025-07-01
- gpt-4o-chatgpt-03-25: released before 2025-07-01
- gpt-4o-mini: released before 2025-07-01
- granite-3-3-8b-instruct: creator outside the providers the catalogue covers
- granite-4-0-350m: creator outside the providers the catalogue covers
- granite-4-0-h-350m: creator outside the providers the catalogue covers
- granite-4-0-h-nano-1b: creator outside the providers the catalogue covers
- granite-4-0-h-small: creator outside the providers the catalogue covers
- granite-4-0-micro: creator outside the providers the catalogue covers
- granite-4-0-nano-1b: creator outside the providers the catalogue covers
- granite-4-1-30b: creator outside the providers the catalogue covers
- granite-4-1-3b: creator outside the providers the catalogue covers
- granite-4-1-8b: creator outside the providers the catalogue covers
- granite-4-2-30b: creator outside the providers the catalogue covers
- granite-4-2-3b: creator outside the providers the catalogue covers
- granite-4-2-8b: creator outside the providers the catalogue covers
- grok-1: released before 2025-07-01
- grok-2-1212: released before 2025-07-01
- grok-3: released before 2025-07-01
- grok-3-mini-reasoning: released before 2025-07-01
- grok-3-reasoning: released before 2025-07-01
- grok-beta: released before 2025-07-01
- hermes-3-llama-3-1-70b: creator outside the providers the catalogue covers
- hermes-4-llama-3-1-405b: creator outside the providers the catalogue covers
- hermes-4-llama-3-1-405b-reasoning: creator outside the providers the catalogue covers
- hermes-4-llama-3-1-70b: creator outside the providers the catalogue covers
- hermes-4-llama-3-1-70b-reasoning: creator outside the providers the catalogue covers
- hy3-non-reasoning: creator outside the providers the catalogue covers
- hyperclova-x-seed-think-32b: creator outside the providers the catalogue covers
- hypernova-60b: creator outside the providers the catalogue covers
- intellect-3: creator outside the providers the catalogue covers
- jamba-1-5-large: creator outside the providers the catalogue covers
- jamba-1-5-mini: creator outside the providers the catalogue covers
- jamba-1-6-large: creator outside the providers the catalogue covers
- jamba-1-6-mini: creator outside the providers the catalogue covers
- jamba-1-7-large: creator outside the providers the catalogue covers
- jamba-1-7-mini: creator outside the providers the catalogue covers
- jamba-reasoning-3b: creator outside the providers the catalogue covers
- jt-35b-flash: creator outside the providers the catalogue covers
- jt-4-1-flash-236b-a21b: creator outside the providers the catalogue covers
- jt-mini: creator outside the providers the catalogue covers
- jt236b: creator outside the providers the catalogue covers
- k-exaone: creator outside the providers the catalogue covers
- k-exaone-2-0-0803: creator outside the providers the catalogue covers
- k-exaone-non-reasoning: creator outside the providers the catalogue covers
- k2-horizon-0-9b: creator outside the providers the catalogue covers
- k2-horizon-3-7b: creator outside the providers the catalogue covers
- k2-horizon-375b-a23b: creator outside the providers the catalogue covers
- k2-horizon-7b: creator outside the providers the catalogue covers
- k2-horizon-mova-36b-a4b: creator outside the providers the catalogue covers
- k2-think-v2: creator outside the providers the catalogue covers
- k2-v2: creator outside the providers the catalogue covers
- k2-v2-low: creator outside the providers the catalogue covers
- k2-v2-medium: creator outside the providers the catalogue covers
- kat-coder-pro-v1: creator outside the providers the catalogue covers
- kat-coder-pro-v2: creator outside the providers the catalogue covers
- lfm-40b: creator outside the providers the catalogue covers
- lfm2-1-2b: creator outside the providers the catalogue covers
- lfm2-2-6b: creator outside the providers the catalogue covers
- lfm2-24b-a2b: creator outside the providers the catalogue covers
- lfm2-5-1-2b-instruct: creator outside the providers the catalogue covers
- lfm2-5-1-2b-thinking: creator outside the providers the catalogue covers
- lfm2-5-2-6b: creator outside the providers the catalogue covers
- lfm2-5-8b-a1b: creator outside the providers the catalogue covers
- lfm2-5-vl-1-6b: creator outside the providers the catalogue covers
- lfm2-8b-a1b: creator outside the providers the catalogue covers
- ling-1t: creator outside the providers the catalogue covers
- ling-2-6-1t: creator outside the providers the catalogue covers
- ling-2-6-flash: creator outside the providers the catalogue covers
- ling-3-0-flash: creator outside the providers the catalogue covers
- ling-3-0-flash-fin: creator outside the providers the catalogue covers
- ling-3-0-flash-vl: creator outside the providers the catalogue covers
- ling-3-0-tiny: creator outside the providers the catalogue covers
- ling-flash-2-0: creator outside the providers the catalogue covers
- ling-mini-2-0: creator outside the providers the catalogue covers
- llama-2-chat-13b: released before 2025-07-01
- llama-2-chat-70b: released before 2025-07-01
- llama-2-chat-7b: released before 2025-07-01
- llama-3-1-instruct-405b: released before 2025-07-01
- llama-3-1-instruct-70b: released before 2025-07-01
- llama-3-1-instruct-8b: released before 2025-07-01
- llama-3-1-nemotron-instruct-70b: creator outside the providers the catalogue covers
- llama-3-1-nemotron-nano-4b-reasoning: creator outside the providers the catalogue covers
- llama-3-1-nemotron-ultra-253b-v1-reasoning: creator outside the providers the catalogue covers
- llama-3-2-instruct-11b-vision: released before 2025-07-01
- llama-3-2-instruct-1b: released before 2025-07-01
- llama-3-2-instruct-3b: released before 2025-07-01
- llama-3-2-instruct-90b-vision: released before 2025-07-01
- llama-3-3-instruct-70b: released before 2025-07-01
- llama-3-3-nemotron-super-49b: creator outside the providers the catalogue covers
- llama-3-3-nemotron-super-49b-reasoning: creator outside the providers the catalogue covers
- llama-3-instruct-70b: released before 2025-07-01
- llama-3-instruct-8b: released before 2025-07-01
- llama-65b: released before 2025-07-01
- llama-nemotron-super-49b-v1-5: creator outside the providers the catalogue covers
- llama-nemotron-super-49b-v1-5-reasoning: creator outside the providers the catalogue covers
- longcat-2-0: creator outside the providers the catalogue covers
- longcat-flash-lite: creator outside the providers the catalogue covers
- magistral-medium: released before 2025-07-01
- magistral-small: released before 2025-07-01
- mercury-2-5: creator outside the providers the catalogue covers
- mi-dm-k-2-5-pro-dec28: creator outside the providers the catalogue covers
- midm-250-pro-rsnsft: creator outside the providers the catalogue covers
- mimo-v2-0206: creator outside the providers the catalogue covers
- mimo-v2-5-0424: creator outside the providers the catalogue covers
- mimo-v2-5-pro: creator outside the providers the catalogue covers
- mimo-v2-5-pro-non-reasoning: creator outside the providers the catalogue covers
- mimo-v2-6-flash: creator outside the providers the catalogue covers
- mimo-v2-6-pro: creator outside the providers the catalogue covers
- mimo-v2-flash: creator outside the providers the catalogue covers
- mimo-v2-flash-reasoning: creator outside the providers the catalogue covers
- mimo-v2-omni: creator outside the providers the catalogue covers
- mimo-v2-omni-0327: creator outside the providers the catalogue covers
- mimo-v2-pro: creator outside the providers the catalogue covers
- minicpm-v4-6-1-3b: creator outside the providers the catalogue covers
- minicpm5-1b: creator outside the providers the catalogue covers
- minicpm5-1b-non-reasoning: creator outside the providers the catalogue covers
- minicpm5-2b: creator outside the providers the catalogue covers
- minimax-m1-40k: creator outside the providers the catalogue covers
- minimax-m1-80k: creator outside the providers the catalogue covers
- minimax-m2-1: creator outside the providers the catalogue covers
- minimax-m2-5: creator outside the providers the catalogue covers
- minimax-m2-7: creator outside the providers the catalogue covers
- mistral-7b-instruct: released before 2025-07-01
- mistral-8x22b-instruct: released before 2025-07-01
- mistral-large: released before 2025-07-01
- mistral-large-2: released before 2025-07-01
- mistral-medium: released before 2025-07-01
- mistral-saba: released before 2025-07-01
- mistral-small: released before 2025-07-01
- mistral-small-2402: released before 2025-07-01
- mistral-small-3: released before 2025-07-01
- mistral-small-3-1: released before 2025-07-01
- mistral-small-3-2: released before 2025-07-01
- mixtral-8x7b-instruct: released before 2025-07-01
- molmo-7b-d: creator outside the providers the catalogue covers
- molmo2-8b: creator outside the providers the catalogue covers
- motif-0714: creator outside the providers the catalogue covers
- motif-2-12-7b: creator outside the providers the catalogue covers
- motif-3: creator outside the providers the catalogue covers
- nanbeige4-1-3b: creator outside the providers the catalogue covers
- nemotron-3-5-lightning: creator outside the providers the catalogue covers
- nemotron-3-nano-omni-30b-a3b: creator outside the providers the catalogue covers
- nemotron-cascade-2-30b-a3b: creator outside the providers the catalogue covers
- nex-n2-pro: creator outside the providers the catalogue covers
- north-mini-code: creator outside the providers the catalogue covers
- nova-2-0-lite: creator outside the providers the catalogue covers
- nova-2-0-lite-reasoning: creator outside the providers the catalogue covers
- nova-2-0-lite-reasoning-low: creator outside the providers the catalogue covers
- nova-2-0-lite-reasoning-medium: creator outside the providers the catalogue covers
- nova-2-0-omni: creator outside the providers the catalogue covers
- nova-2-0-omni-reasoning-low: creator outside the providers the catalogue covers
- nova-2-0-omni-reasoning-medium: creator outside the providers the catalogue covers
- nova-2-0-pro: creator outside the providers the catalogue covers
- nova-2-0-pro-reasoning-low: creator outside the providers the catalogue covers
- nova-2-0-pro-reasoning-medium: creator outside the providers the catalogue covers
- nova-lite: creator outside the providers the catalogue covers
- nova-micro: creator outside the providers the catalogue covers
- nova-premier: creator outside the providers the catalogue covers
- nova-pro: creator outside the providers the catalogue covers
- nvidia-nemotron-3-nano-30b-a3b: creator outside the providers the catalogue covers
- nvidia-nemotron-3-nano-30b-a3b-reasoning: creator outside the providers the catalogue covers
- nvidia-nemotron-3-nano-4b: creator outside the providers the catalogue covers
- nvidia-nemotron-3-super-120b-a12b: creator outside the providers the catalogue covers
- nvidia-nemotron-3-ultra-550b-a55b: creator outside the providers the catalogue covers
- nvidia-nemotron-nano-12b-v2-vl: creator outside the providers the catalogue covers
- nvidia-nemotron-nano-12b-v2-vl-reasoning: creator outside the providers the catalogue covers
- nvidia-nemotron-nano-9b-v2: creator outside the providers the catalogue covers
- nvidia-nemotron-nano-9b-v2-reasoning: creator outside the providers the catalogue covers
- o1: released before 2025-07-01
- o1-mini: released before 2025-07-01
- o1-preview: released before 2025-07-01
- o1-pro: released before 2025-07-01
- o3: released before 2025-07-01
- o3 Mini: released before 2025-07-01
- o3 Mini High: released before 2025-07-01
- o3 Pro: released before 2025-07-01
- o4 Mini: released before 2025-07-01
- olmo-2-32b: creator outside the providers the catalogue covers
- olmo-2-7b: creator outside the providers the catalogue covers
- olmo-3-1-32b-instruct: creator outside the providers the catalogue covers
- olmo-3-1-32b-think: creator outside the providers the catalogue covers
- olmo-3-32b-think: creator outside the providers the catalogue covers
- olmo-3-7b-instruct: creator outside the providers the catalogue covers
- olmo-3-7b-think: creator outside the providers the catalogue covers
- openchat-35: creator outside the providers the catalogue covers
- palm-2: released before 2025-07-01
- phi-3-mini: creator outside the providers the catalogue covers
- phi-4-mini: creator outside the providers the catalogue covers
- phi-4-multimodal: creator outside the providers the catalogue covers
- pixtral-large-2411: released before 2025-07-01
- quasar-438b: creator outside the providers the catalogue covers
- qwen-2-5-max: released before 2025-07-01
- qwen-chat-14b: released before 2025-07-01
- qwen-chat-72b: released before 2025-07-01
- qwen-turbo: released before 2025-07-01
- qwen1.5-110b-chat: released before 2025-07-01
- qwen2-5-72b-instruct: released before 2025-07-01
- qwen2-5-coder-32b-instruct: released before 2025-07-01
- qwen2-5-coder-7b-instruct: released before 2025-07-01
- qwen2-72b-instruct: released before 2025-07-01
- qwen2.5-32b-instruct: released before 2025-07-01
- qwen3-0.6b-instruct: released before 2025-07-01
- qwen3-0.6b-instruct-reasoning: released before 2025-07-01
- qwen3-1.7b-instruct: released before 2025-07-01
- qwen3-1.7b-instruct-reasoning: released before 2025-07-01
- qwen3-14b-instruct: released before 2025-07-01
- qwen3-14b-instruct-reasoning: released before 2025-07-01
- qwen3-235b-a22b-instruct: released before 2025-07-01
- qwen3-235b-a22b-instruct-reasoning: released before 2025-07-01
- qwen3-30b-a3b-instruct: released before 2025-07-01
- qwen3-30b-a3b-instruct-reasoning: released before 2025-07-01
- qwen3-32b-instruct: released before 2025-07-01
- qwen3-32b-instruct-reasoning: released before 2025-07-01
- qwen3-4b-instruct: released before 2025-07-01
- qwen3-4b-instruct-reasoning: released before 2025-07-01
- qwen3-8b-instruct: released before 2025-07-01
- qwen3-8b-instruct-reasoning: released before 2025-07-01
- qwq-32b: released before 2025-07-01
- r1-1776: creator outside the providers the catalogue covers
- reka-flash: creator outside the providers the catalogue covers
- ring-1t: creator outside the providers the catalogue covers
- ring-2-6-1t: creator outside the providers the catalogue covers
- ring-flash-2-0: creator outside the providers the catalogue covers
- sarvam-105b: creator outside the providers the catalogue covers
- sarvam-30b: creator outside the providers the catalogue covers
- sarvam-m-reasoning: creator outside the providers the catalogue covers
- seed-oss-36b-instruct: creator outside the providers the catalogue covers
- solar-mini: creator outside the providers the catalogue covers
- solar-mini4: creator outside the providers the catalogue covers
- solar-open-100b-reasoning: creator outside the providers the catalogue covers
- solar-open2-250b: creator outside the providers the catalogue covers
- solar-pro-2: creator outside the providers the catalogue covers
- solar-pro-2-preview: creator outside the providers the catalogue covers
- solar-pro-2-preview-reasoning: creator outside the providers the catalogue covers
- solar-pro-2-reasoning: creator outside the providers the catalogue covers
- sonar: creator outside the providers the catalogue covers
- sonar-pro: creator outside the providers the catalogue covers
- sonar-reasoning: creator outside the providers the catalogue covers
- sonar-reasoning-pro: creator outside the providers the catalogue covers
- step-3-5-flash: creator outside the providers the catalogue covers
- step-3-5-flash-0202: creator outside the providers the catalogue covers
- step-3-7-flash: creator outside the providers the catalogue covers
- step-3-vl-10b: creator outside the providers the catalogue covers
- step-5: creator outside the providers the catalogue covers
- tiny-aya-global: creator outside the providers the catalogue covers
- tri-21b-think-preview: creator outside the providers the catalogue covers
- tri-21b-think-v0-5: creator outside the providers the catalogue covers
- tulu3-405b: creator outside the providers the catalogue covers
Data sourced from Artificial Analysis. Licence: Artificial Analysis commercial data licence.