HP - Unigine: Heaven

System:
GPU: Nvidia 940MX 4GB [1241MHz]
CPU: Intel i7-7500U [2.9GHz*] [2 Phys. Cores]
* - The average speed returned by the benchmark. Base clock @ 2.7GHz, Turbo @ 3.5GHz

Settings:

API: 				OpenGL [1]; DX11 [2]
Quality: 			High
Tess. 				Normal
Stereo3D: 			Disabled
Multimonitor: 		Disabled
AA: 				x2
Fullscreen: 		Yes
Resolution: 		System [1920x1080]
===================================================================

Results [OpenGL]:

Avg. FPS: 			10.9
Min. FPS: 			4.4
Max. FPS: 			22.5
Score: 				275
===================================================================

Results [DX11]:

Avg. FPS: 			10.6
Min. FPS: 			6.9
Max. FPS: 			20.6
Score: 				268
===================================================================

NOTES:

This GPU does not support DX12.

Average temperatures are not shown in the benchmark results. But I would imagine they are somewhat important, so I watched the benchmark periodically to see what temperatures it fluctuated between.
For OpenGL, it was around 75C - 79C.
For DX11, it was at a lower 72C - 74C.

I have thermal throttling 'disabled' in Nvidia Control Panel (Rather, I have it set to 'prefer consistent performance').
Thus, MAX FPS may be slightly lower in these tests than if this was set to 'prefer maximum performance' although that would decrease the Avg. FPS by creating massive frame drops for seconds at a time.

OpenGL achieved a higher score, so I'll write future benchmarks (and this quick comparison) in terms of its performance relative to DX11, rather than vice-versa.

Score:		275 	vs. 	268  (+2.6%)
Avg. FPS:	10.1 	vs.		10.6 (-4.8%)
Min. FPS:	4.4		vs.		6.9  (-36.2%)
Max. FPS:	22.5	vs.		20.6 (+9.2%)

OpenGL has the edge to pump out more impressive numbers, but only at the cost of instability and higher temperatures.
Seems that DX11 is able to achieve better consistency, at a small sacrifice of performing 2.6% worse.