benchmark
.
space
benchmarks
rankings
compare
voices
transcripts
papers
articles
power rankings
Who's winning
Composite score: average rank percentile across benchmarks + breadth bonus. Models with 3+ benchmarks. Higher = consistently ranks near the top across more benchmarks.
1
Claude Mythos 5
Anthropic · 6 benchmarks
92.1
2
Claude Mythos Preview
Anthropic · 13 benchmarks
91.9
3
GPT-5.6 Sol Ultra
OpenAI · 3 benchmarks
90.6
4
GPT-5.6 Terra
OpenAI · 9 benchmarks
86.8
5
Claude Fable 5
Anthropic · 30 benchmarks
86.2
6
Qwen 3.6 Plus
Alibaba · 9 benchmarks
84.8
7
Kimi K3
Moonshot AI · 28 benchmarks
84.1
8
GPT-5.5 Pro
OpenAI · 7 benchmarks
82.0
9
Bonsai 27B (Ternary)
PrismML · 8 benchmarks
80.5
10
Qwen3.7-Max
Alibaba · 8 benchmarks
77.9
11
GPT-5.4 Pro
OpenAI · 8 benchmarks
77.3
12
Surfer 2
H Company · 4 benchmarks
77.0
13
OpenVLThinkerV2
arxiv · 5 benchmarks
75.8
14
Qwen 3.5 397B
Alibaba · 10 benchmarks
74.8
15
Gemini 3 Pro
Google · 5 benchmarks
74.7
16
DeepSeek V3.2 Speciale
DeepSeek · 3 benchmarks
73.5
17
Seed 2.0 Pro
ByteDance · 6 benchmarks
73.2
18
GPT-5.6 Sol
OpenAI · 35 benchmarks
72.2
19
Kimi K2.6
Moonshot AI · 15 benchmarks
71.7
20
GPT-5.3 Codex
OpenAI · 6 benchmarks
71.3
21
DeepSeek V3.2
DeepSeek · 7 benchmarks
68.9
22
Claude Opus 4.7
Anthropic · 28 benchmarks
68.9
23
GPT-5.6 Luna
OpenAI · 9 benchmarks
68.5
24
GPT-5.2
OpenAI · 6 benchmarks
68.4
25
GLM-4.7
Zhipu AI · 5 benchmarks
67.7
26
OpenVLThinkerV2 8B
Research · 6 benchmarks
67.6
27
DeepSeek V4 Pro
DeepSeek · 16 benchmarks
67.3
28
Gemini 3.1 Pro
Google · 30 benchmarks
67.2
29
Arcee Trinity
Arcee AI · 6 benchmarks
67.2
30
Gemini 3.5 Flash
Google DeepMind · 11 benchmarks
67.0