Skip to main content
BenchLM
Data

APEX-Agents-AA

We show this table for reference; we do not rank on it.

Data verified

Artificial Analysis' implementation of the APEX-Agents benchmark for long-horizon professional-services agent tasks.

Benchmark score on APEX-Agents-AA — November 13, 2026

We compile the APEX-Agents-AA rows from secondary reports. Gemini 3.5 Flash leads the table at 47.1%, followed by Qwen3.8 Max Preview (42.4%) and Kimi K3 (41.3%). We do not use these results to rank models overall.

1Closed

Gemini 3.5 Flash

Google

Carried from an upstream page
47.1%
Overall 61.98Context 1M
2Closed

Qwen3.8 Max Preview

Alibaba

Carried from an upstream page
42.4%
3Open weights

Kimi K3

Moonshot AI

Carried from an upstream page
41.3%
Overall 68.42Context 1.05M

30 modelsAgenticCurrentDisplay onlyUpdated November 13, 2026

Benchmark score table (30 models)

None of the 30 results on this page has a documented run by the benchmark or an independent evaluator: 30 carried from an upstream page, where who ran them is not recorded.

Benchmark score results for APEX-Agents-AA
RankModel / configurationScoreParameters (B)Open / closed
1Gemini 3.5 FlashGoogleCarried from an upstream page
47.1%
Not reportedClosed
2Qwen3.8 Max PreviewAlibabaCarried from an upstream page
42.4%
Not reportedClosed
3Kimi K3Moonshot AICarried from an upstream page
41.3%
Not reportedOpen
4GPT-5.6 TerraOpenAICarried from an upstream page
38.9%
Not reportedClosed
5Step 5 PreviewStepFunCarried from an upstream page
38.0%
Not reportedOpen
6GPT-5.5OpenAICarried from an upstream page
37.7%
Not reportedClosed
7GPT-5.6 LunaOpenAICarried from an upstream page
35.8%
Not reportedClosed
8GLM-5.2Z.AICarried from an upstream page
33.7%
Not reportedOpen
9GPT-5.4OpenAICarried from an upstream page
33.3%
Not reportedClosed
10Claude Opus 4.6 (Adaptive)AnthropicCarried from an upstream page
33.0%
Not reportedClosed
11Gemini 3.1 ProGoogleCarried from an upstream page
32.0%
Not reportedClosed
12Apodex 1.1ApodexCarried from an upstream page
31.2%
Not reportedClosed
13Apodex 1.1 MiniApodexCarried from an upstream page
31.2%
Not reportedOpen
14Kimi K2.6Moonshot AICarried from an upstream page
28.5%
Not reportedOpen
15GPT-5.4 miniOpenAICarried from an upstream page
28.2%
Not reportedClosed
16Ling 3.0 Flash FinInclusionAICarried from an upstream page
27.4%
Not reportedClosed
17GPT-5.4 nanoOpenAICarried from an upstream page
24.9%
Not reportedClosed
18DeepSeek V4 Pro 0813DeepSeekCarried from an upstream page
24.3%
Not reportedOpen
19Qwen3.7 PlusAlibabaCarried from an upstream page
22.4%
Not reportedClosed
20Grok 4.3xAICarried from an upstream page
17.0%
Not reportedClosed
21Qwen3.5 397BAlibabaCarried from an upstream page
15.3%
Not reportedOpen
22Qwen3.5 397B (Reasoning)AlibabaCarried from an upstream page
15.3%
Not reportedOpen
23Step 3.7 FlashStepFunCarried from an upstream page
14.8%
Not reportedOpen
24GLM-5Z.AICarried from an upstream page
14.5%
Not reportedOpen
25Kimi K2.5 (Reasoning)Moonshot AICarried from an upstream page
11.5%
Not reportedOpen
26Kimi K2.5Moonshot AICarried from an upstream page
11.5%
Not reportedOpen
27MiniMax M2.7MiniMaxCarried from an upstream page
10.6%
Not reportedOpen
28GPT-OSS 120BOpenAICarried from an upstream page
3.1%
Not reportedOpen
29MiMo-V2.5-ProXiaomiCarried from an upstream page
2.4%
Not reportedClosed
30GPT-OSS 20BOpenAICarried from an upstream page
0.7%
Not reportedOpen
  • Lab-reported
  • Benchmark-run
  • Independent
What the labels mean

Among the reported APEX-Agents-AA rows, Gemini 3.5 Flash is first at 47.1%. The third row is 5.8 points behind. The broader top-10 range is 14.1 points, so the table still separates the published systems.

30 models have been evaluated on APEX-Agents-AA. The benchmark falls in the Agentic category. APEX-Agents-AA is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About APEX-Agents-AA

Year

2026

Tasks

452 professional-services agent tasks

Format

Pass@1

Difficulty

Long-horizon workplace agent tasks

BenchLM stores APEX-Agents-AA as a display-only agentic row. Artificial Analysis reports pass@1 over 452 public APEX-Agents tasks spanning investment banking, management consulting, and corporate law.

APEX-Agents-AA Benchmark Leaderboard

Freshness and provenance

Version

APEX-Agents-AA 2026

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

CurrentDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

Questions

What does APEX-Agents-AA measure?

Artificial Analysis' implementation of the APEX-Agents benchmark for long-horizon professional-services agent tasks.

Which model scores highest on APEX-Agents-AA?

Gemini 3.5 Flash by Google currently leads with a score of 47.1% on APEX-Agents-AA.

How many models are evaluated on APEX-Agents-AA?

30 AI models have published results on APEX-Agents-AA in the BenchLM catalog.

Compare top models on APEX-Agents-AA

Gemini 3.5 Flash vs Qwen3.8 Max PreviewQwen3.8 Max Preview vs Kimi K3Kimi K3 vs GPT-5.6 TerraGPT-5.6 Terra vs Step 5 Preview
Last updated: November 13, 2026 · BenchLM version APEX-Agents-AA 2026

Know when it’s worth switching models

The model to choose, the cheaper alternative, and the release we would wait on.

Read a sample issue

Join 7,000+ subscribers.

One email each week. Unsubscribe anytime.