Spatial Reasoning

LiveBench · 28 models

Rank	Model	Score
1	claude-fable-5-xhigh-effort	100
1	claude-opus-4-7-xhigh-effort	100
1	claude-sonnet-4-6-thinking-auto-medium-effort	100
1	claude-sonnet-5-xhigh-effort	100
1	qwen3.6-27b	100
6	claude-opus-4-6-thinking-auto-high-effort	98
6	claude-opus-4-8-xhigh-effort	98
6	gemini-3.1-pro-preview-high	98
6	gpt-5.4-xhigh	98
6	gpt-5.5	98
6	gpt-5.5-high	98
6	gpt-5.5-xhigh	98
6	grok-4.3	98
6	grok-build-0.1	98
6	qwen3.6-plus	98
16	claude-opus-4-5-20251101-thinking-64k-high-effort	96
16	deepseek-v4-flash	96
16	gemini-3.5-flash-high	96
16	gpt-5.2-2025-12-11-high	96
16	qwen3.7-max	96
21	deepseek-v4-pro	94
21	glm-5.2	94
21	gpt-5.2-codex	94
21	kimi-k2.6-thinking	94
21	minimax-m3	94
26	kimi-k2.7-code	92
27	gpt-5.4-mini-xhigh	88
28	gpt-5.4-nano-xhigh	82