NVIDIA · Maxwell
NVIDIA GeForce GTX 850M: specs, benchmarks and comparison
- 40 W
- Release
- March 12, 2014
- Power
- 40 W
- depends on the laptop model
Performance
Rank among models tested in the same benchmark on the same render backend. A laptop chip’s result depends on the laptop model: its maker sets the power limit and cooling.
Blender Benchmark 2.93.1
GPU, bmw27 (CUDA)
412.99 s
lower is better
Median of 1 runs · small sample · Blender Open Data (Blender Foundation) · November 14, 2026
Other benchmark versions (4)
| Benchmark | Result | Rank |
|---|---|---|
| Blender Benchmark 2.90 | 518.05 s · small sample | — |
| Blender Benchmark 2.83 | 498.4 s · small sample | — |
| Blender Benchmark 2.82 | 488.9 s · small sample | — |
| Blender Benchmark 2.81 | 871.25 s · small sample | — |
Each benchmark version has its own scale; results from different versions are not compared.
Blender Benchmark 2.83
GPU, classroom (CUDA)
1,907.46 s
lower is better
Median of 1 runs · small sample · Blender Open Data (Blender Foundation) · November 14, 2026
Other benchmark versions (1)
| Benchmark | Result | Rank |
|---|---|---|
| Blender Benchmark 2.81 | 3,941.18 s · small sample | — |
Each benchmark version has its own scale; results from different versions are not compared.
Blender Benchmark 2.83
GPU, fishy_cat (CUDA)
1,003.87 s
lower is better
Median of 1 runs · small sample · Blender Open Data (Blender Foundation) · November 14, 2026
Blender Benchmark 2.90
GPU, koro (CUDA)
1,782.88 s
lower is better
Median of 1 runs · small sample · Blender Open Data (Blender Foundation) · November 14, 2026
Other benchmark versions (1)
| Benchmark | Result | Rank |
|---|---|---|
| Blender Benchmark 2.83 | 1,999.4 s · small sample | — |
Each benchmark version has its own scale; results from different versions are not compared.
Blender Benchmark 2.83
GPU, pavillon_barcelona (CUDA)
2,802.18 s
lower is better
Median of 1 runs · small sample · Blender Open Data (Blender Foundation) · November 14, 2026
Blender Benchmark 4.5.0
GPU, sum of scenes monster + junkshop + classroom (CUDA)
143 samples/min
higher is better
Median of 1 runs · small sample · Blender Open Data (Blender Foundation) · November 14, 2026
Other benchmark versions (4)
| Benchmark | Result | Rank |
|---|---|---|
| Blender Benchmark 4.4.0 | 106 samples/min · small sample | — |
| Blender Benchmark 3.4.0 | 119 samples/min · small sample | — |
| Blender Benchmark 3.2.1 | 136 samples/min · small sample | — |
| Blender Benchmark 3.1.0 | 120 samples/min · small sample | — |
Each benchmark version has its own scale; results from different versions are not compared.
Blender Benchmark 4.5.0
GPU, sum of scenes monster + junkshop + classroom (OptiX)
102 samples/min
higher is better
Median of 2 runs · small sample · Blender Open Data (Blender Foundation) · November 14, 2026
Other benchmark versions (5)
| Benchmark | Result | Rank |
|---|---|---|
| Blender Benchmark 4.4.0 | 94 samples/min · small sample | — |
| Blender Benchmark 4.3.0 | 102 samples/min · small sample | — |
| Blender Benchmark 4.1.0 | 97 samples/min · small sample | — |
| Blender Benchmark 4.0.0 | 95 samples/min · small sample | — |
| Blender Benchmark 3.4.0 | 111 samples/min · small sample | — |
Each benchmark version has its own scale; results from different versions are not compared.
Specifications
Chip maker values. The laptop maker sets the power limit, clocks and memory in each laptop.
Chip
- GPU chip
- GM107
- Architecture
- Maxwell
Memory
- Memory bus
- 128 bit
Power and dimensions
- Board power
- 40 W (TDP)
- Interface
- PCIe 3.0 x16
Several configurations exist under this name. The table lists the shared specifications. Clocks and power limit depend on the laptop model.
Specifications: Wikipedia (English), checked November 14, 2026. Release date: Wikipedia (CC BY-SA 4.0).