← LiveBench leaderboards

AMPS Hard

LiveBench · 28 models

RankModelScore
1claude-fable-5-xhigh-effort99
1claude-opus-4-5-20251101-thinking-64k-high-effort99
3claude-opus-4-7-xhigh-effort98
3claude-opus-4-8-xhigh-effort98
3claude-sonnet-5-xhigh-effort98
3deepseek-v4-flash98
3deepseek-v4-pro98
3gemini-3.1-pro-preview-high98
3gemini-3.5-flash-high98
3glm-5.298
3gpt-5.2-codex98
3gpt-5.4-nano-xhigh98
3gpt-5.4-xhigh98
3gpt-5.598
3gpt-5.5-high98
3qwen3.7-max98
17claude-opus-4-6-thinking-auto-high-effort97
17gpt-5.2-2025-12-11-high97
17gpt-5.4-mini-xhigh97
17gpt-5.5-xhigh97
17grok-4.397
17kimi-k2.6-thinking97
17kimi-k2.7-code97
17qwen3.6-plus97
25qwen3.6-27b93
26claude-sonnet-4-6-thinking-auto-medium-effort76
27minimax-m375
28grok-build-0.166