AA Progress Index v1
As of 28 Oct 2026, the top score on AA Progress Index v1 is 351.3 by Claude Opus 5.5 (Anthropic), ahead of Claude Sonnet 5.5 at 350.5, across 630 model configurations with a published result (Artificial Analysis, result dated 22 Sept 2026). Artificial Analysis' Progress Index v1: a long-run item-response capability scale fitted across many evaluations so model scores stay comparable as individual tests saturate. Anchored so o3-mini lands near 220 and GPT-5.5 near 320. AA Progress Index v1 is shown for reference only and does not count toward the BenchLeader Index.
Key facts (as of 28 Oct 2026)
- Claude Opus 5.5 holds the top AA Progress Index v1 score at 351.3 (Artificial Analysis, 22 Sept 2026).
- Claude Sonnet 5.5 is second on AA Progress Index v1 at 350.5 (Artificial Analysis, 28 Sept 2026).
- AA Progress Index v1: Artificial Analysis' long-run IRT capability scale (~200–350) fitted across many evaluations so scores stay comparable as individual tests saturate. Shown for trends; not part of the BenchLeader Index (Artificial Analysis).
- 630 model configurations have a published AA Progress Index v1 result on BenchLeader (BenchLeader, 28 Oct 2026).
- AA Progress Index v1 does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).
Artificial Analysis' long-run IRT capability scale (~200–350) fitted across many evaluations so scores stay comparable as individual tests saturate. Shown for trends; not part of the BenchLeader Index.
- Published by
- Artificial Analysis
- Category
- Composite
- Index weight
- Reference only
- Models
- 630
- Data as of
- 28 Oct 2026
- Newest result
- 26 Oct 2026
Source: Artificial Analysis (artificialanalysis.ai). Data taken from the public leaderboard.
What the test looks like
Artificial Analysis' Progress Index v1: a long-run item-response capability scale fitted across many evaluations so model scores stay comparable as individual tests saturate. Anchored so o3-mini lands near 220 and GPT-5.5 near 320.
How it is scored
An IRT-style score typically between about 200 and 350; higher is better. Artificial Analysis fits each effort variant separately. Shown for trends only, not part of the BenchLeader Index.
What to keep in mind
A model-based estimate rather than a single-test measurement, and refits can move every score at once. Prefer the Intelligence Index and the BenchLeader Index for current rankings; use this for long-run progress charts.
- 1Claude Opus 5.5 (xhigh)351.3
- 2Claude Sonnet 5.5 (max)350.5
- 3Claude Opus 5.5 (max)350
- 4GPT-6 Astra (max)349.6
- 5GPT-6.1 Sol (max)348.5
- 6GPT-6 Astra (xhigh)348.4
- 7GPT-6 Astra (high)347.4
- 8GPT-6.1 Sol (xhigh)346.4
- 9GPT-6.1 Terra (max)345.2
- 10Claude Opus 5.5 (high)344
- 11GPT-6 Astra (medium)344
- 12GPT-6.1 Sol (high)343.9
- 13Claude Sonnet 5.5 (xhigh)343.8
- 14Muse Wonder Full (max)343.2
- 15Gemini 4 Argon (high)341.8
| # | ||||
|---|---|---|---|---|
| 1 |  Claude Opus 5.5xhigh | 351.3 | 66.8 | 2026-09-22 |
| 2 | Claude Sonnet 5.5max | 350.5 | 65.7 | 2026-09-28 |
| 3 | Claude Opus 5.5max | 350 | 69.8 | 2026-09-22 |
| 4 |  GPT-6 Astramax | 349.6 | 68.3 | 2026-09-03 |
| 5 | GPT-6.1 Solmax | 348.5 | 67.6 | 2026-09-29 |
| 6 | GPT-6 Astraxhigh | 348.4 | 67.1 | 2026-09-03 |
| 7 | GPT-6 Astrahigh | 347.4 | 67.3 | 2026-09-03 |
| 8 | GPT-6.1 Solxhigh | 346.4 | 64.8 | 2026-09-29 |
| 9 | NewGPT-6.1 Terramax | 345.2 | 65.0 | 2026-10-26 |
| 10 | Claude Opus 5.5high | 344 | 65.9 | 2026-09-22 |
| 11 | GPT-6 Astramedium | 344 | 65.2 | 2026-09-03 |
| 12 | GPT-6.1 Solhigh | 343.9 | 64.3 | 2026-09-29 |
| 13 | Claude Sonnet 5.5xhigh | 343.8 | 63.4 | 2026-09-28 |
| 14 |  Muse Wonder Fullmax | 343.2 | 62.7 | 2026-10-12 |
| 15 |  Gemini 4 Argonhigh | 341.8 | 66.0 | 2026-09-30 |
| 16 | Claude Fable 5.1xhigh | 340.9 | 65.0 | 2026-09-01 |
| 17 | Claude Fable 5.1max | 340.8 | 66.9 | 2026-09-01 |
| 18 | New Kimi K3.1 Pro | 340.7 | 59.8 | 2026-10-23 |
| 19 | Claude Opus 5.5medium | 340.3 | 62.2 | 2026-09-22 |
| 20 | GPT-6.1 Solmedium | 338.7 | 63.1 | 2026-09-29 |
| 21 | Claude Fable 5.1high | 337.9 | 63.5 | 2026-09-01 |
| 22 | NewGPT-6.1 Terraxhigh | 337.9 | 60.4 | 2026-10-26 |
| 23 | Muse Wondermax | 335.8 | 61.2 | 2026-10-12 |
| 24 | GPT-6 Astralow | 335.1 | 62.0 | 2026-09-03 |
| 25 | NewGPT-6.1 Terrahigh | 334.1 | 59.9 | 2026-10-26 |
| 26 | GPT-6 Solmax | 334 | 63.6 | 2026-09-22 |
| 27 | Claude Opus 5max | 333.7 | 65.5 | 2026-07-24 |
| 28 | Claude Sonnet 5.5high | 333.5 | 57.9 | 2026-09-28 |
| 29 | Claude Fable 5.1medium | 333 | 61.5 | 2026-09-01 |
| 30 | Claude Opus 5xhigh | 333 | 60.3 | 2026-07-24 |
| 31 | GPT-5.6 Solmax | 331.5 | 65.4 | 2026-07-09 |
| 32 | Claude Opus 5high | 331 | 64.4 | 2026-07-24 |
| 33 | Claude Fable 5max | 329.3 | 65.3 | 2026-06-09 |
| 34 | NewGPT-6.1 Terramedium | 328.9 | 58.3 | 2026-10-26 |
| 35 | Claude Fable 5.1low | 328.4 | 58.1 | 2026-09-01 |
| 36 | GPT-6 Solxhigh | 327.3 | 58.8 | 2026-09-22 |
| 37 | GPT-6.1 Sollow | 326.2 | 59.0 | 2026-09-29 |
| 38 | Claude Haiku 5.5max | 326.1 | 56.4 | 2026-10-07 |
| 39 | Claude Opus 5.5low | 325 | 57.8 | 2026-09-22 |
| 40 | Claude Opus 5medium | 324.9 | 57.5 | 2026-07-24 |
| 41 | Muse Spark 1.3max | 324.5 | 60.9 | 2026-09-02 |
| 42 | GPT-5.6 Solxhigh | 324.3 | 64.4 | 2026-07-09 |
| 43 | GPT-6 Solhigh | 324.1 | 58.4 | 2026-09-22 |
| 44 | GPT-5.6 Terramax | 323.9 | 61.2 | 2026-07-09 |
| 45 |  Qwen3.8 Maxmax | 323.7 | 59.1 | 2026-09-02 |
| 46 |  DeepSeek V4.1 Flash 1017maxopen ↗ | 323.7 | 57.8 | 2026-10-17 |
| 47 |  GLM 5.3maxopen ↗ | 322.7 | 57.6 | 2026-08-18 |
| 48 |  Grok 4.7xhigh | 322.7 | 57.4 | 2026-09-21 |
| 49 | Claude Haiku 5.5xhigh | 322.2 | 56.0 | 2026-10-07 |
| 50 | Grok 4.7high | 321.6 | 56.1 | 2026-09-21 |
| 51 |  Ling 3.1 Flashopen | 321.6 | 49.8 | 2026-10-22 |
| 52 | Grok 4.6high | 321.5 | 60.6 | 2026-08-12 |
| 53 | GPT-5.6 Solhigh | 321.2 | 63.6 | 2026-07-09 |
| 54 |  MiMo-V2.6-Proopen ↗ | 321.2 | 56.9 | 2026-09-21 |
| 55 | Grok 4.6xhigh | 320.8 | 58.5 | 2026-08-12 |
| 56 | Gemini 3.8 Flashhigh | 320.3 | 60.3 | 2026-09-02 |
| 57 | DeepSeek V4.1 Flashmaxopen ↗ | 320.2 | 58.9 | 2026-09-10 |
| 58 | Qwen3.8-Flash-Nextopen ↗ | 320.2 | 54.7 | 2026-08-26 |
| 59 | GPT-5.5xhigh | 320 | 64.4 | 2026-04-23 |
| 60 | Kimi K3maxopen ↗ | 319.9 | 62.2 | 2026-07-16 |
More composite benchmarks · All LLM benchmarks
Cite as: BenchLeader, “AA Progress Index v1 leaderboard”, data as of 28 Oct 2026, https://www.benchleader.com/benchmarks/aa_progress_index.
AA Progress Index v1: questions
- What does AA Progress Index v1 measure?
- Artificial Analysis' Progress Index v1: a long-run item-response capability scale fitted across many evaluations so model scores stay comparable as individual tests saturate. Anchored so o3-mini lands near 220 and GPT-5.5 near 320. Scores are reported in score; higher is better.
- What is the top score on AA Progress Index v1?
- As of 28 Oct 2026, the highest score on AA Progress Index v1 is 351.3 by Claude Opus 5.5 (result dated 22 Sept 2026), ahead of Claude Sonnet 5.5 at 350.5. Scores come from Artificial Analysis and are re-read daily.
- Which AI model leads AA Progress Index v1?
- Claude Opus 5.5 leads AA Progress Index v1 with 351.3 as of 28 Oct 2026, ahead of Claude Sonnet 5.5 at 350.5 (result dated 22 Sept 2026).
- How many models have AA Progress Index v1 results?
- 630 model configurations have a AA Progress Index v1 result on BenchLeader, all taken from Artificial Analysis.
- Who runs AA Progress Index v1 and how often is it updated?
- AA Progress Index v1 is published by Artificial Analysis. BenchLeader re-reads the published results once a day and records the date each result was published.
- Does AA Progress Index v1 count toward the BenchLeader Index?
- No. AA Progress Index v1 is shown for reference but left out of the composite index.