Best models for professional workflows

Agents’ Last Exam

Evaluates completion of long professional workflows with graded deliverables. This track uses the 122-task v1.1 release from October 17, without the October 26 update.

Updated: Nov 12, 2026Data as of: Nov 12, 2026

Limited reportingNot saturated

Official benchmark

Agents’ Last Exam results. Each row identifies its benchmark version and evaluation track. Scores from different versions are not directly comparable. Best effort from the latest stored report date per model and version.
Model / creator	Version	Reported score
Gemini 4 HydrogenGoogle · Proprietary	1.1 / October 17 task set / Google Hydrogen launch	71.6%
GPT-6.1 TerraOpenAI · Proprietary	1.1 / October 17 task set / Google Hydrogen launch	69.8%
Claude Sonnet 5.5Anthropic · Proprietary	1.1 / October 17 task set / Google Hydrogen launch	60.4%

View all 3 published reports

Gemini 4 Hydrogen71.6%

1.1 / October 17 task set / Google Hydrogen launch

effort: highest available or best reported · Hydrogen: Antigravity CLI 1.4.8, 5h, safety filters. Sonnet: ALE-Claw under the reported setup. Terra: public board. Harnesses differ. · Tools enabled

Google DeepMind ↗

GPT-6.1 Terra69.8%

1.1 / October 17 task set / Google Hydrogen launch

effort: highest available or best reported · Hydrogen: Antigravity CLI 1.4.8, 5h, safety filters. Sonnet: ALE-Claw under the reported setup. Terra: public board. Harnesses differ. · Tools enabled

Google DeepMind ↗

Claude Sonnet 5.560.4%

1.1 / October 17 task set / Google Hydrogen launch

effort: highest available or best reported · Hydrogen: Antigravity CLI 1.4.8, 5h, safety filters. Sonnet: ALE-Claw under the reported setup. Terra: public board. Harnesses differ. · Tools enabled

Google DeepMind ↗

Sources and original reports ↗

Other benchmarks

LiveBench: Coding

Tests coding tasks, helping you compare models for implementation work that you would otherwise write or debug yourself.

Vision2Web

Tests building functional websites from visual prototypes and requirements, helping you compare models for turning a client's design into working code.

DeepFinance

Tests private equity and investment banking work such as valuation, due diligence and financial modeling, helping you assess AI support for demanding client analysis.