ai.canshu.com
中文
VRAM Calculator Advisor & TCO API Pricing Compare Models Leaderboard GPUs

Search models, GPUs, and tools

↑↓ Navigate · Enter Open · ★ Watchlist prioritized ai.canshu.com

LLM Leaderboard & Multi-Dimension Comparison

Synced 2026-11-20

154 models × 375 public benchmarks across 18 capability dimensions. · Top CN: Kimi K3.1 Pro (2.8T MoE) (#10)

Overall #1
Anthropic Claude Fable 5.2
63.7
Coding #1
Anthropic Claude Fable 5.2
55.0
Reasoning #1
Anthropic Claude Fable 5.2
63.7
Agents #1
Google DeepMind Gemini 4 Neon
51.1
Core
Modality
Language
Domains

🏆 Overall Leaderboard154 models

Score = TrueSkill μ aggregated over all public benchmarks in this dimension

1
Anthropic
Claude Fable 5.2
ClosedAnthropic
63.7
28 benchmarks
2
Google DeepMind
Gemini 4 Neon
ClosedGoogle DeepMind
63.2
11 benchmarks
3
OpenAI
GPT-6.2 Astra
ClosedOpenAI
62.4
12 benchmarks
#4
Anthropic
Claude Opus 5.5
ClosedAnthropic
60.8
31 benchmarks
#5
Google DeepMind
Gemini 4 Carbon
ClosedGoogle DeepMind
60.7
15 benchmarks
#6
Anthropic
Claude Sonnet 5.5
ClosedAnthropic
58.9
41 benchmarks
#7
OpenAI
GPT-6 Astra
ClosedOpenAI
58.6
22 benchmarks
#8
OpenAI
GPT-6.2 Sol
ClosedOpenAI
57.7
12 benchmarks
#9
Anthropic
Claude Fable 5.1
ClosedAnthropic
57.1
6 benchmarks
#10
Moonshot Kimi
Kimi K3.1 Pro (2.8T MoE)
OpenMoonshot Kimi
56.7
30 benchmarks
#11
Google DeepMind
Gemini 4 Argon
ClosedGoogle DeepMind
56.4
18 benchmarks
#12
Anthropic
Claude Opus 5
ClosedAnthropic
56.1
12 benchmarks
#13
OpenAI
GPT-6.1 Sol
ClosedOpenAI
55.5
16 benchmarks
#14
Baidu 百度文心
ERNIE-5.5-Flash (280B MoE)
OpenBaidu 百度文心
55.2
13 benchmarks
#15
Meta
Muse Spark 1.3
ClosedMeta
55.1
10 benchmarks
#16
Anthropic
Claude Fable 5
ClosedAnthropic
54.8
19 benchmarks
#17
Alibaba Qwen
Qwen4-Max / Qwen3.8-Max
ClosedAlibaba Qwen
54.0
11 benchmarks
#18
OpenAI
GPT-5.6 Sol
ClosedOpenAI
53.9
45 benchmarks
#19
Google DeepMind
Gemini 4 Hydrogen
ClosedGoogle DeepMind
53.8
5 benchmarks
#20
53.7
11 benchmarks
#21
Moonshot Kimi
Kimi K3 (2.8T MoE)
OpenMoonshot Kimi
52.9
29 benchmarks
#22
Zhipu AI
GLM-5.3 (753B MoE)
OpenZhipu AI
52.5
17 benchmarks
#23
Meituan 美团 LongCat
Meituan LongCat-2.5 (1.78T MoE)
OpenMeituan 美团 LongCat
52.2
19 benchmarks
#24
Meta
Muse Wonder
ClosedMeta
52.2
16 benchmarks
#25
OpenAI
GPT-5.5 Pro
ClosedOpenAI
51.6
5 benchmarks
#26
Google DeepMind
Gemini 3.8 Flash
ClosedGoogle DeepMind
51.5
13 benchmarks
#27
OpenAI
GPT-6 Sol
ClosedOpenAI
51.4
5 benchmarks
#28
Anthropic
Claude Haiku 5.5
ClosedAnthropic
51.4
34 benchmarks
#29
Anthropic
Claude Opus 4.8
ClosedAnthropic
51.4
27 benchmarks
#30
Meta
Muse Spark 1.1
ClosedMeta
51.4
11 benchmarks

Source: ZeroEval / llm-stats.com public API (vendor-reported + third-party), synced daily (last 2026-11-20). Harness settings vary by vendor.