APEX-Agents (AA) leaderboard
As of 6 Nov 2026, the top score on APEX-Agents (AA) is 47.0% by Gemini 3.5 Flash (Google), ahead of Qwen3.8 Max at 42.4%, across 33 model configurations with a published result (Artificial Analysis, result dated 19 May 2026). APEX-Agents sets long-horizon professional tasks written and graded by domain experts at Mercor. APEX-Agents (AA) is shown for reference only and does not count toward the BenchLeader Index.
Key facts (as of 6 Nov 2026)
- Gemini 3.5 Flash holds the top APEX-Agents (AA) score at 47.0% (Artificial Analysis, 19 May 2026).
- Qwen3.8 Max is second on APEX-Agents (AA) at 42.4% (Artificial Analysis, 2 Sep 2026).
- APEX-Agents (AA): Artificial Analysis' run of Mercor's long-horizon professional agent tasks (Artificial Analysis).
- 33 model configurations have a published APEX-Agents (AA) result on BenchLeader (BenchLeader, 6 Nov 2026).
- APEX-Agents (AA) does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).
Artificial Analysis' run of Mercor's long-horizon professional agent tasks.
- Published by
- Artificial Analysis
- Category
- Agents & tools
- Index weight
- Reference only
- Models
- 33
- Data as of
- 6 Nov 2026
Source: Artificial Analysis (artificialanalysis.ai). Data taken from the public leaderboard.
What the test looks like
APEX-Agents sets long-horizon professional tasks written and graded by domain experts at Mercor.
How it is scored
Percent of tasks judged successful, run by Artificial Analysis.
What to keep in mind
The Epoch-mirrored official results are already shown; this is an independent re-run.
- 1Gemini 3.5 Flash (high)47.0%
- 2Qwen3.8 Max (max)42.4%
- 3Kimi K3 (max)41.3%
- 4GPT-5.6 Terra (max)38.9%
- 5Step 5 Preview38.0%
- 6GPT-5.5 (xhigh)37.7%
- 7GPT-5.6 Luna (max)35.8%
- 8GLM 5.2 (max)33.7%
- 9GPT-5.4 (xhigh)33.3%
- 10Claude Opus 4.6 (max)33.0%
- 11Gemini 3.1 Pro32.0%
- 12Apodex 1.131.2%
- 13Kimi K2.6 (thinking)28.5%
- 14GPT-5.4 Mini (xhigh)28.2%
- 15Claude Sonnet 4.6 (max)28.0%
| # | ||||
|---|---|---|---|---|
| 1 |  Gemini 3.5 Flashhigh | 47.0% | 62.5 | 2026-05-19 |
| 2 |  Qwen3.8 Maxmax | 42.4% | 58.8 | 2026-09-02 |
| 3 |  Kimi K3maxopen ↗ | 41.3% | 62.1 | 2026-07-16 |
| 4 |  GPT-5.6 Terramax | 38.9% | 60.7 | 2026-07-09 |
| 5 |  Step 5 Previewopen | 38.0% | 54.9 | 2026-09-18 |
| 6 | GPT-5.5xhigh | 37.7% | 64.3 | 2026-04-23 |
| 7 | GPT-5.6 Lunamax | 35.8% | 56.3 | 2026-07-09 |
| 8 |  GLM 5.2maxopen ↗ | 33.7% | 59.5 | 2026-06-16 |
| 9 | GPT-5.4xhigh | 33.3% | 63.2 | 2026-03-05 |
| 10 |  Claude Opus 4.6max | 33.0% | 55.8 | 2026-02-05 |
| 11 | Gemini 3.1 Pro | 32.0% | 60.5 | 2026-02-19 |
| 12 | Apodex 1.1 | 31.2% | – | 2026-08-30 |
| 13 | Kimi K2.6thinkingopen ↗ | 28.5% | 59.7 | 2026-04-20 |
| 14 | GPT-5.4 Minixhigh | 28.2% | 56.0 | 2026-03-17 |
| 15 | Claude Sonnet 4.6max | 28.0% | 54.2 | 2026-02-17 |
| 16 | Gemini 3 Flashthinking | 27.7% | 58.9 | 2025-12-17 |
| 17 |  Ling 3.0 Flash Finopen | 27.4% | – | 2026-09-11 |
| 18 | GPT-5.4 Nanoxhigh | 24.9% | 54.0 | 2026-03-17 |
| 19 |  DeepSeek V4 Pro 0424maxopen | 24.3% | – | 2026-04-24 |
| 20 | Qwen3.7 Plus | 22.4% | 57.7 | 2026-06-01 |
| 21 |  Grok 4.3high | 17.0% | 56.6 | 2026-04-30 |
| 22 | Qwen3.5 397B A17Bthinkingopen ↗ | 15.3% | 56.6 | 2026-02-16 |
| 23 | Step 3.7 Flashopen ↗ | 14.8% | 53.0 | 2026-05-29 |
| 24 | DeepSeek V3.2thinkingopen ↗ | 14.5% | 56.6 | 2025-12-01 |
| 25 | GLM-5thinkingopen ↗ | 14.4% | 60.2 | 2026-02-11 |
| 26 | Grok 4.20thinking | 14.2% | 58.4 | 2026-03-10 |
| 27 | Gemini 3.1 Flash Lite | 12.2% | 52.4 | 2026-03-03 |
| 28 | Kimi K2.5thinkingopen ↗ | 11.5% | 58.7 | 2026-01-27 |
| 29 |  MiniMax M2.7open ↗ | 10.6% | 56.5 | 2026-03-18 |
| 30 | gpt-oss-120bhighopen ↗ | 3.1% | 48.9 | 2025-08-05 |
| 31 |  MiMo-V2.5-Prothinkingopen ↗ | 2.4% | – | 2026-04-22 |
| 32 |  Nemotron 3 Super 120B A12Bthinkingopen ↗ | 1.8% | 51.1 | 2026-03-11 |
| 33 | gpt-oss-20bhighopen ↗ | 0.7% | 45.5 | 2025-08-05 |
More agents & tools benchmarks · All LLM benchmarks
Related benchmarks
Top head-to-heads on APEX-Agents (AA)
Cite as: BenchLeader, “APEX-Agents (AA) leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/aa_apex_agents.
APEX-Agents (AA): questions
- What does APEX-Agents (AA) measure?
- APEX-Agents sets long-horizon professional tasks written and graded by domain experts at Mercor. Scores are reported in percent of tasks solved; higher is better.
- What is the top score on APEX-Agents (AA)?
- As of 6 Nov 2026, the highest score on APEX-Agents (AA) is 47.0% by Gemini 3.5 Flash (result dated 19 May 2026), ahead of Qwen3.8 Max at 42.4%. Scores come from Artificial Analysis and are re-read daily.
- Which AI model leads APEX-Agents (AA)?
- Gemini 3.5 Flash leads APEX-Agents (AA) with 47.0% as of 6 Nov 2026 (result dated 19 May 2026), ahead of Qwen3.8 Max at 42.4%.
- How many models have APEX-Agents (AA) results?
- 33 model configurations have a APEX-Agents (AA) result on BenchLeader, all taken from Artificial Analysis.
- Who runs APEX-Agents (AA) and how often is it updated?
- APEX-Agents (AA) is published by Artificial Analysis. BenchLeader re-reads the published results once a day and records the date each result was published.
- Does APEX-Agents (AA) count toward the BenchLeader Index?
- No. APEX-Agents (AA) is shown for reference but left out of the composite index.