BenchLeader
Menu
Best LLMsChoose a modelLeaderboardsModelsBenchmarksProvidersReleases9

AA Intelligence Index v4.3.2

Artificial Analysis' composite of its own evaluations (GPQA, HLE, IFBench, LCR, τ²-Bench, Terminal-Bench Hard, SciCode, AA-Omniscience and more).

As of 25 Oct 2026, Claude Opus 5.5 leads AA Intelligence Index v4.3.2 on BenchLeader with 57.6, ahead of Claude Sonnet 5.5 at 56, across 633 model configurations with a published result.

Published by
Artificial Analysis
Category
Composite
Index weight
1.0
Models
633
Data as of
25 Oct 2026
Newest result
23 Oct 2026

Source: Artificial Analysis (artificialanalysis.ai). Data taken from the public leaderboard.

What the test looks like

Artificial Analysis' composite of the evaluations it runs itself, currently including GPQA Diamond, Humanity's Last Exam, IFBench, AA-LCR, τ²-Bench Telecom, Terminal-Bench Hard, SciCode and AA-Omniscience, each run in a fixed harness.

How it is scored

An equally weighted average of the component scores, on Artificial Analysis' 0 to 100 scale. Shown for reference; BenchLeader uses the components rather than the composite.

What to keep in mind

Its membership changes between versions, which shifts every score at once. It is one publisher's view of what matters; the BenchLeader Index draws on many.

Top 15
  1. 1Claude Opus 5.5 (max)57.6
  2. 2Claude Sonnet 5.5 (max)56
  3. 3Claude Opus 5.5 (xhigh)56.0
  4. 4Muse Wonder Full (max)55.8
  5. 5Kimi K3.1 Pro55.1
  6. 6Claude Opus 5.5 (high)53.6
  7. 7Claude Fable 5.1 (max)53.4
  8. 8Claude Fable 5.1 (xhigh)53.2
  9. 9GPT-6 Astra (max)52.7
  10. 10Gemini 4 Argon (high)52.6
  11. 11GPT-6 Astra (xhigh)52.4
  12. 12Muse Wonder (max)52.4
  13. 13Claude Sonnet 5.5 (xhigh)51.9
  14. 14GPT-6.1 Sol (max)51.8
  15. 15Claude Opus 5.5 (medium)51.2
633 of 633
#
1Claude Opus 5.5max57.672.62026-09-22
2Claude Sonnet 5.5max5669.42026-09-28
3Claude Opus 5.5xhigh56.070.02026-09-22
4Muse Wonder Fullmax55.870.62026-10-12
5NewKimi K3.1 Pro55.166.92026-10-23
6Claude Opus 5.5high53.669.42026-09-22
7Claude Fable 5.1max53.470.52026-09-01
8Claude Fable 5.1xhigh53.269.52026-09-01
9GPT-6 Astramax52.770.42026-09-03
10Gemini 4 Argonhigh52.670.72026-09-30
11GPT-6 Astraxhigh52.468.92026-09-03
12Muse Wondermax52.467.52026-10-12
13Claude Sonnet 5.5xhigh51.967.92026-09-28
14GPT-6.1 Solmax51.869.92026-09-29
15Claude Opus 5.5medium51.266.52026-09-22
16Claude Fable 5.1high51.169.42026-09-01
17GPT-6.1 Solxhigh51.067.72026-09-29
18GPT-6 Astrahigh50.969.62026-09-03
19Claude Opus 5max50.868.82026-07-24
20GPT-6.1 Solhigh50.267.72026-09-29
21Claude Opus 5xhigh49.767.42026-07-24
22Claude Fable 5max49.669.32026-06-09
23GPT-6 Astramedium49.667.52026-09-03
24Claude Fable 5.1medium48.967.32026-09-01
25Claude Opus 5high48.167.92026-07-24
26Muse Spark 1.3max48.167.12026-09-02
27GPT-6.1 Solmedium47.867.22026-09-29
28GPT-6 Solmax47.666.42026-09-22
29GPT-5.6 Solmax47.068.02026-07-09
30Claude Fable 5.1low46.865.02026-09-01
31Claude Sonnet 5.5high46.865.62026-09-28
32Grok 4.7xhigh46.561.62026-09-21
33Grok 4.7high46.363.92026-09-21
34MiMo-V2.6-Pro46.364.02026-09-21
35GPT-6 Astralow45.865.92026-09-03
36Qwen3.8 Maxmax45.463.32026-09-02
37DeepSeek V4.1 Flash 1017max45.164.02026-10-17
38Muse Spark 1.3xhigh45.164.92026-09-02
39Claude Opus 5medium44.864.62026-07-24
40GLM 5.3max44.863.52026-08-18
41Grok 4.6high44.362.92026-08-12
42GPT-6 Solxhigh44.264.32026-09-22
43Grok 4.6xhigh44.262.52026-08-12
44GPT-5.6 Solxhigh44.066.62026-07-09
45Ling 3.1 Flash43.9–2026-10-22
46Step 5 Preview43.760.52026-09-18
47Kimi K3max43.664.82026-07-16
48Claude Haiku 5.5max43.460.52026-10-07
49Grok 4.6medium42.864.22026-08-12
50NewMiniMax M3.1 Flash42.761.92026-10-19
51GPT-6 Solhigh42.463.72026-09-22
52GPT-5.6 Solhigh42.466.32026-07-09
53Claude Opus 5.5low42.364.72026-09-22
54Grok 4.7low42.261.02026-09-21
55GPT-6.1 Sollow42.164.92026-09-29
56GPT-5.6 Terramax42.164.42026-07-09
57GLM 5.3 Flash41.862.72026-08-26
58Claude Opus 4.8max41.863.32026-05-28
59Claude Haiku 5.5xhigh41.361.52026-10-07
60Gemini 3.8 Flashhigh40.964.42026-09-02

More composite benchmarks · All LLM benchmarks

Cite as: BenchLeader, “AA Intelligence Index v4.3.2 leaderboard”, data as of 25 Oct 2026, https://www.benchleader.com/benchmarks/aa_intelligence_index.

AA Intelligence Index v4.3.2: questions

What does AA Intelligence Index v4.3.2 measure?
Artificial Analysis' composite of the evaluations it runs itself, currently including GPQA Diamond, Humanity's Last Exam, IFBench, AA-LCR, τ²-Bench Telecom, Terminal-Bench Hard, SciCode and AA-Omniscience, each run in a fixed harness. Scores are reported in index; higher is better.
Which AI model leads AA Intelligence Index v4.3.2?
Claude Opus 5.5 leads AA Intelligence Index v4.3.2 with 57.6 as of 25 Oct 2026, ahead of Claude Sonnet 5.5 at 56.
How many models have AA Intelligence Index v4.3.2 results?
633 model configurations have a AA Intelligence Index v4.3.2 result on BenchLeader, all taken from Artificial Analysis.
Who runs AA Intelligence Index v4.3.2 and how often is it updated?
AA Intelligence Index v4.3.2 is published by Artificial Analysis. BenchLeader re-reads the published results once a day and records the date each result was published.
Does AA Intelligence Index v4.3.2 count toward the BenchLeader Index?
Yes. AA Intelligence Index v4.3.2 contributes to the composite category of the BenchLeader Index, normalised so that 50 is the average of the evaluated models.