Rank01Proprietary
Claude Fable 5.1
Anthropic
- Reasoning index #1
- Text preference #3
- Agent tasks #1
98
100%
3 benchmark signals
LLM Leaderboard
A traceable LLM ranking snapshot that brings together multiple public benchmark leaderboards into one decision table, refreshed every day.
Models
20
Signals
44
Updated
2026-09-08
Refreshed daily at 00:00 HKT
Consensus snapshot
Major benchmark signals, normalized into one decision table.
#1Claude Fable 5.1Anthropic · Reasoning index #1 · Text preference #3 · Agent tasks #198
#2Claude Fable 5Anthropic · Reasoning index #4 · Text preference #1 · Agent tasks #393
#3GPT-6 AstraOpenAI · Reasoning index #290
#4Claude Opus 5Anthropic · Reasoning index #3 · Text preference #7 · Agent tasks #288
Consensus Ranking
The consensus score is not a raw benchmark score. It normalizes cross-benchmark positions and groups effort levels and variants of one product line under its best result.
Weight mixReasoning index 40%Text preference 30%Agent tasks 30%
RankModelScoreCoverage
Rank02Proprietary
Claude Fable 5
Anthropic
- Reasoning index #4
- Text preference #1
- Agent tasks #3
93
100%
3 benchmark signals
Rank03Proprietary
GPT-6 Astra
OpenAI
- Reasoning index #2
90
40%
1 benchmark signals
Rank04Proprietary
Claude Opus 5
Anthropic
- Reasoning index #3
- Text preference #7
- Agent tasks #2
88
100%
3 benchmark signals
Rank05Proprietary
Muse Spark 1.3
Meta
- Reasoning index #5
79
40%
1 benchmark signals
Rank06Proprietary
GPT-5.6 Sol
OpenAI
- Reasoning index #6
- Text preference #14
- Agent tasks #4
73
100%
3 benchmark signals
Rank07Proprietary
Claude Opus 4.7
Anthropic
- Text preference #4
- Agent tasks #10
73
60%
2 benchmark signals
Rank08Open weights
Kimi K3
Kimi
- Reasoning index #10
- Text preference #10
- Agent tasks #6
69
100%
3 benchmark signals
Rank09Proprietary
Claude Opus 4.6
Anthropic
- Text preference #2
- Agent tasks #17
63
60%
2 benchmark signals
Rank10Open weights
Hy4 preview
Tencent
- Agent tasks #9
63
30%
1 benchmark signals
Rank11Proprietary
Claude Opus 4.8
Anthropic
- Text preference #15
- Agent tasks #5
61
60%
2 benchmark signals
Rank12Proprietary
Gemini 3.8 Flash
- Reasoning index #14
- Text preference #6
- Agent tasks #14
58
100%
3 benchmark signals
Rank13Proprietary
Muse Spark
Meta
- Text preference #11
56
30%
1 benchmark signals
Rank14Proprietary
Qwen3.8 Max
Alibaba
- Reasoning index #7
- Text preference #19
- Agent tasks #13
54
100%
3 benchmark signals
Rank15Proprietary
GPT 5.5
OpenAI
- Text preference #16
- Agent tasks #8
54
60%
2 benchmark signals
Rank16Open weights
GLM-5.3
Z AI
- Reasoning index #8
- Text preference #17
- Agent tasks #18
49
100%
3 benchmark signals
Rank17Proprietary
Gemini 3 Pro
- Text preference #13
48
30%
1 benchmark signals
Rank18Proprietary
Muse Spark 1.2
Meta
- Reasoning index #16
- Text preference #5
- Agent tasks #28
41
100%
3 benchmark signals
Rank19Open weights
Qwen3.8 2.4T A95B
Alibaba
- Reasoning index #15
41
40%
1 benchmark signals
Rank20Proprietary
Grok 4.6
SpaceXAI
- Reasoning index #9
- Text preference #42
- Agent tasks #16
39
100%
3 benchmark signals
Methodology
Weighted consensus, not benchmark shopping.
- 01Three public benchmark leaderboards are fetched automatically every day at 00:00 HKT and cached for 24 hours.
- 02Each benchmark rank is converted into position points, then averaged with the declared weight mix.
- 03Effort levels and variants of one product line are grouped, keeping only its best result per benchmark.
- 04Models missing some benchmark coverage receive a small coverage discount, so a single leaderboard spike does not dominate the table.
- 05The ranking is a procurement and experimentation entry point, not an absolute answer. Cost, latency, context, data policy, and task-level testing still matter.
Turn rankings into a practical model selection process.
If you want to connect this leaderboard to model selection, prompt tests, or agent workflows, we can turn it into a repeatable evaluation lane.