APEX-Agents: Investment Banking Analyst Created by experts from Goldman Sachs, Morgan Stanley, JPMorgan, Barclays
Show sample task for Gemini 4 Argon Gemini 4 ArgonHigh
81.9%
Show sample task for Sonnet 5.5 Sonnet 5.5Max
2 efforts77.2%
Show sample task for Sonnet 5.5 Show sample task for Sonnet 5.5 Show sample task for Gemini 3.7 Flash Gemini 3.7 FlashHigh
71.3%
Show sample task for Opus 5.5 Opus 5.5Max
2 efforts69.3%
Show sample task for Opus 5.5 Show sample task for Opus 5.5 Show sample task for Opus 5 Opus 5Max
66.6%
Show sample task for Grok 4.6 Grok 4.6xHigh
65.0%
Show sample task for Qwen3.8-Max Qwen3.8-MaxxHigh
64.2%
Show sample task for Fable 5 Fable 5Max
64.1%
Show sample task for Fable 5.1 Fable 5.1Max
2 efforts62.5%
Show sample task for Fable 5.1 Show sample task for Fable 5.1 Show sample task for GLM-5.3 GLM-5.3Max
60.6%
Show sample task for GPT-6.1 Sol GPT-6.1 SolMax
58.1%
Show sample task for Haiku 5.5 Haiku 5.5Max
57.5%
Show sample task for MiMo V2.6 Flash RL MiMo V2.6 Flash RLAuto
57.2%
Show sample task for DeepSeek-V4.1-Flash-1017 DeepSeek-V4.1-Flash-1017Max
57.2%
Show sample task for GPT-6 Astra GPT-6 AstraMax
55.9%
Show sample task for DeepSeek-V4-Flash DeepSeek-V4-FlashMax
55.4%
Show sample task for GPT-5.6 Terra GPT-5.6 TerraMax
55.3%
Show sample task for Opus 4.8 Opus 4.8Max
55.3%
Show sample task for GPT-6 Sol GPT-6 SolMax
55.2%
Show sample task for GPT-5.4 GPT-5.4xHigh
55.0%
Show sample task for Grok 4.5 Grok 4.5High
54.7%
Show sample task for Sonnet 5 Sonnet 5Max
54.1%
Show sample task for GPT-5.5 GPT-5.5xHigh
52.8%
Show sample task for Muse Spark 1.3 Muse Spark 1.3xHigh
2 efforts51.6%
Show sample task for Muse Spark 1.3 Show sample task for Muse Spark 1.3 Show sample task for MiMo V2.6 Pro MiMo V2.6 ProAuto
50.0%
Show sample task for Gemini 3.8 Flash Gemini 3.8 FlashHigh
49.7%
Show sample task for GLM-5.3-Flash GLM-5.3-FlashMax
49.7%
Show sample task for GPT-5.6 Sol GPT-5.6 SolMax • Pro
49.7%
Show sample task for Opus 4.7 Opus 4.7Max
48.1%
Show sample task for Beam BeamMax
48.1%
Show sample task for Grok 4.7 Grok 4.7xHigh
47.8%
Show sample task for GLM-5.2 GLM-5.2Max
46.5%
Show sample task for Mercury 3 Mercury 3High
46.2%
Show sample task for Gemini 3.6 Flash Gemini 3.6 FlashHigh
45.6%
Show sample task for Kimi K3 Kimi K3Max
44.1%
Show sample task for Qwen 3.8 27B Qwen 3.8 27BxHigh
43.8%
Show sample task for Sonnet 4.6 Sonnet 4.6High
43.4%
Show sample task for DeepSeek-V4-Pro-0813 DeepSeek-V4-Pro-0813Max
43.1%
Show sample task for Opus 4.6 Opus 4.6Max
40.5%
Show sample task for GLM-5.1 GLM-5.1
39.6%
Show sample task for GPT-6 Luna GPT-6 LunaMax
38.8%
Show sample task for DeepSeek-V4.1-Flash DeepSeek-V4.1-FlashMax
38.8%
Show sample task for Gemini 3.1 Pro Gemini 3.1 ProHigh
37.5%
Show sample task for MiniMax-M3 MiniMax-M3High
36.9%
Show sample task for Muse Spark 1.2 Muse Spark 1.2xHigh
36.2%
Show sample task for GPT-5.6 Luna GPT-5.6 LunaMax
35.3%
Show sample task for Kimi K2.7 Code Kimi K2.7 CodeHigh
34.2%
Show sample task for Gemini 3.5 Flash Lite Gemini 3.5 Flash LiteHigh
29.7%
Show sample task for Inkling InklingHigh
26.3%
Show sample task for Muse Spark Open Muse Spark OpenxHigh
19.7%
Show sample task for Qwen3.5 Qwen3.5
19.1%
Show sample task for Muse Spark 1.1 Muse Spark 1.1xHigh
16.0%
Show sample task for DeepSeek-V3.2 DeepSeek-V3.2
13.9%
Show sample task for Nemotron 3 Ultra Nemotron 3 UltraHigh
13.6%
Show sample task for Gemini 3.5 Flash Gemini 3.5 FlashHigh
11.9%
Show sample task for GPT-OSS-120B GPT-OSS-120BHigh
0.9%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Sample task (1 out of 80) Task
Task
Environment
Environment
Trajectory
Trajectory
This sample task is provided for illustration only. The scores may not reflect an agent's overall performance on the leaderboard.
Assess how much certain business drivers must move to bring IRR below 20%. The stock price has fallen to $87.00. Use the LBO model for the info.
Find the critical point (percentages to 2 decimal places) for each of the below business drivers at which rounded IRR would be pushed down to 19.99% from above:
a) 'Growth rate scale' (correct the approach for Year 1/2)
b) 'Customer acquisition costs'
c) 'R&D cost'
d) 'Debt costs'
e) 'EBITDA multiple'
Assumptions and constraints:
EBITDA is not to fall below $91 million for any year. If this threshold is passed the new constraint becomes EBITDA for each individual year instead of the IRR.
'Debt costs' should be set to 0% for all other sensitivities.
Create a new sheet that shows values for the five major business drivers.
No rubric data available for this model.