APEX-Agents-AA

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

Data verified September 27, 202634 confirmed releases in the last 30 daysFollow model changes

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

Benchmark score on APEX-Agents-AA — September 27, 2026

We compile the APEX-Agents-AA rows from secondary reports. Gemini 3.5 Flash leads the table at 47.1%, followed by Kimi K3 (41.3%) and GPT-5.6 Terra (38.9%). We do not use these results to rank models overall.

1Proprietary

Gemini 3.5 Flash

Google

47.1%

Overall 63.49Context 1M

2Proprietary

Kimi K3

Moonshot AI

41.3%

Overall 71.87Context 1.05M

3Proprietary

GPT-5.6 Terra

OpenAI

38.9%

Overall 72.58Context 1.05M

27 modelsAgenticCurrentDisplay onlyUpdated September 27, 2026

Benchmark score table (27 models)

Score

Gemini 3.5 FlashGoogle · Closed

47.1%

Kimi K3Moonshot AI · Closed

41.3%

GPT-5.6 TerraOpenAI · Closed

38.9%

GPT-5.5OpenAI · Closed

37.7%

GPT-5.6 LunaOpenAI · Closed

35.8%

GLM-5.2Z.AI · Open weight

33.7%

GPT-5.4OpenAI · Closed

33.3%

Claude Opus 4.6 (Adaptive)Anthropic · Closed

33.0%

Gemini 3.1 ProGoogle · Closed

32.0%

Apodex 1.1Apodex · Closed

31.2%

Apodex 1.1 MiniApodex · Open weight

31.2%

Kimi K2.6Moonshot AI · Open weight

28.5%

GPT-5.4 miniOpenAI · Closed

28.2%

GPT-5.4 nanoOpenAI · Closed

24.9%

DeepSeek V4 Pro 0813DeepSeek · Open weight

24.3%

Qwen3.7 PlusAlibaba · Closed

22.4%

Grok 4.3xAI · Closed

17.0%

Step 3.7 FlashStepFun · Open weight

14.8%

GLM-5Z.AI · Open weight

14.5%

Kimi K2.5Moonshot AI · Open weight

11.5%

Kimi K2.5 (Reasoning)Moonshot AI · Closed

11.5%

MiniMax M2.7MiniMax · Open weight

10.6%

GPT-OSS 120BOpenAI · Open weight

3.1%

MiMo-V2.5-ProXiaomi · Closed

2.4%

Nemotron 3 Super 100BNVIDIA · Open weight

1.8%

Nemotron 3 Super 120B A12BNVIDIA · Open weight

1.8%

GPT-OSS 20BOpenAI · Open weight

0.7%

Among the reported APEX-Agents-AA rows, Gemini 3.5 Flash is first at 47.1%. The third row is 8.2 points behind. The broader top-10 range is 15.9 points, so the table still separates the published systems.

27 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?

27 AI models have been evaluated on APEX-Agents-AA on BenchLM.

Compare top models on APEX-Agents-AA

Gemini 3.5 Flash vs Kimi K3 Kimi K3 vs GPT-5.6 Terra GPT-5.6 Terra vs GPT-5.5 GPT-5.5 vs GPT-5.6 Luna

Last updated: September 27, 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 2,000+ readers.

One email each week. Unsubscribe anytime.