Intel Arc Pro B60 vs NVIDIA GeForce GT 745M Comparison

FAQ

Q: What is the performance gap between the NVIDIA GeForce GT 745M and the Intel Arc Pro B60?

A: The GT 745M has an average benchmark score of 3953, while the Arc Pro B60 averages 3182. The GT 745M is roughly 24% higher in average score, though the Arc Pro B60’s nearest rivals include the GeForce GT 640 at 3210 and the GeForce 920M at 3287, indicating its score sits in a similar low-to-mid range.

Q: Which GPU has more memory and bandwidth?

A: The Intel Arc Pro B60 offers 24 GB of GDDR6 memory on a 192-bit bus with 456.0 GB/s bandwidth. The GT 745M has 2 GB of GDDR5 on a 128-bit bus with 64.00 GB/s bandwidth. The Arc Pro B60 provides 7.1 times the bandwidth of the older part.

Q: What are the key architectural differences?

A: The GT 745M uses NVIDIA’s Kepler architecture on a 28 nm TSMC process with 1,270 million transistors. The Arc Pro B60 is built on Intel’s Xe2-HPG architecture (Battlemage Pro Series) on a 5 nm TSMC node with 19,600 million transistors, making it roughly 15.4 times denser in transistor count per area.

Q: Which GPU supports newer APIs?

A: The Arc Pro B60 supports DirectX 12 Ultimate (12_2), Vulkan 1.4, and OpenGL 4.6. The GT 745M supports DirectX 12 (11_0), Vulkan 1.2.175, and OpenGL 4.6. The Intel part has a higher DirectX feature level and a newer Vulkan version.

Q: How do their physical sizes and power requirements compare?

A: The GT 745M is an integrated graphics product (IGP) with a 45 W TDP and no power connectors. The Arc Pro B60 is a dual-slot card measuring 167 mm in length, with a 200 W TDP, a single 8-pin power connector, and a suggested PSU of 550 W.

Q: What is the release status of each GPU?

A: The GT 745M is end-of-life, released on 2013-03-31, with its successor being the GeForce 800M. The Arc Pro B60 is active, released on 2025-09-04, and has no predecessor or successor listed.

Architecture Differences

The two GPUs come from different design eras and philosophies. The GT 745M is built on Kepler, NVIDIA’s 2012-era architecture, using TSMC’s 28 nm process. It packs 1,270 million transistors into a 118 mm² die, yielding a transistor density of 10.8 million per square millimeter. The chip is GK107 and belongs to the GeForce 700M generation.

The Arc Pro B60 uses Intel’s Xe2-HPG architecture from the Battlemage Pro Series. It is fabricated on a 5 nm process at TSMC, with 19,600 million transistors on a 272 mm² die. That gives a transistor density of 72.1 million per square millimeter, roughly 6.7 times higher than the Kepler part. This density advantage enables far more compute resources: 2560 shading units, 160 TMUs, and 80 ROPs versus 384 shading units, 32 TMUs, and 16 ROPs on the GT 745M.

The Arc Pro B60 also includes 20 dedicated ray tracing cores, which the GT 745M lacks entirely. Neither GPU lists tensor cores. Peak FP32 throughput is 12.29 TFLOPS on the Arc versus 421.6 GFLOPS on the GT 745M, a 29-fold difference. The Intel card also supports FP16 at 24.58 TFLOPS via a 2:1 ratio, while the older NVIDIA part has no listed FP16 rate.

Clocks differ significantly. The Arc Pro B60 has a base clock of 2000 MHz and a boost of 2400 MHz. The GT 745M does not list base or boost clocks, only a memory clock of 1000 MHz with 4 Gbps effective data rate. The Arc’s memory runs at 2375 MHz with 19 Gbps effective throughput, over a 192-bit bus. Combined with 24 GB of GDDR6, the memory configuration is far more robust for large data sets.

Pixel and texture fill rates follow the same hierarchy: the Arc Pro B60 outputs 192.0 GPixel/s and 384.0 GTexel/s, while the GT 745M manages 4.392 GPixel/s and 17.57 GTexel/s. That is roughly 44 times the pixel throughput and 22 times the texture throughput, a generational gap that appears in every measured workload.

The interface changes as well. The GT 745M uses PCIe 3.0 x16, while the Arc Pro B60 uses PCIe 5.0 x8, offering double the per-lane bandwidth of the older standard. Display outputs for the GT 745M are portable device dependent, whereas the Arc Pro B60 includes four mini-DisplayPort 2.1 outputs.

---

The Verdict

The data shows a clear split by use case. The GT 745M is an older, mobile-oriented integrated GPU with a respectable average score of 3953. Its nearest rivals are all low-power parts: the AMD Radeon R5 M420 at 3956, the NVIDIA GeForce 830M at 3957, and the NVIDIA Quadro K2000 at 3964. The GT 745M sits within a 0.6% delta of these. For a 2013-era laptop GPU, it remains competitive within its niche.

The Arc Pro B60 is a different class of hardware. Despite a lower average score of 3182, its architecture and memory system target professional workloads, not raw legacy benchmark averages. Its average score is close to the NVIDIA Quadro P1000 at 3163, and it is 3.5% above the RTX 5080 SUPER’s 3075 average, though the RTX 5080 SUPER is not a direct performance peer in most modern tests.

Who should pick the GT 745M? Anyone needing a legacy, low-power integrated GPU for a portable device, provided the workload is limited to older DirectX 11 titles or basic OpenCL tasks. Its 23rd percentile ranking among all GPUs indicates it is outclassed by nearly three-quarters of the field.

Who should pick the Arc Pro B60? The 20th percentile suggests a low overall standing, but the 24 GB memory and 456 GB/s bandwidth make it a candidate for memory-hungry applications like AI inference or large dataset rendering. The 12.29 TFLOPS FP32 throughput and DP display outputs point to a professional workstation card, not a gaming part. The 200 W TDP and 550 W PSU requirement confirm it is meant for a desktop chassis with proper cooling.

For gaming, the GT 745M may still handle older titles at low settings, but the Arc Pro B60’s API support (DirectX 12 Ultimate, Vulkan 1.4) makes it more future-proof. However, the GT 745M’s higher average benchmark score suggests that in pure legacy DirectX 11 tests, the Kepler part remains competitive, which is a notable result given the 12-year gap between releases.

---

Specification Differences

| Specification | NVIDIA GeForce GT 745M | Intel Arc Pro B60 |

|---------------|------------------------|-------------------|

| Architecture | Kepler | Xe2-HPG |

| Generation | GeForce 700M | Battlemage (Pro Series) |

| Process Node | 28 nm | 5 nm |

| Transistors | 1,270 million | 19,600 million |

| Die Size | 118 mm² | 272 mm² |

| Transistor Density | 10.8M / mm² | 72.1M / mm² |

| Memory Size | 2 GB | 24 GB |

| Memory Type | GDDR5 | GDDR6 |

| Memory Bus | 128 bit | 192 bit |

| Memory Bandwidth | 64.00 GB/s | 456.0 GB/s |

| Shading Units | 384 | 2560 |

| TMUs | 32 | 160 |

| ROPs | 16 | 80 |

| RT Cores | 0 | 20 |

| Pixel Rate | 4.392 GPixel/s | 192.0 GPixel/s |

| Texture Rate | 17.57 GTexel/s | 384.0 GTexel/s |

| FP32 | 421.6 GFLOPS | 12.29 TFLOPS |

| FP16 | N/A | 24.58 TFLOPS (2:1) |

| TDP | 45 W | 200 W |

| Slot Width | IGP | Dual-slot |

| Power Connectors | None | 1x 8-pin |

| Bus Interface | PCIe 3.0 x16 | PCIe 5.0 x8 |

| Display Outputs | Portable Device Dependent | 4x mini-DisplayPort 2.1 |

| DirectX Support | 12 (11_0) | 12 Ultimate (12_2) |

| Vulkan Support | 1.2.175 | 1.4 |

| Release Date | 2013-03-31 | 2025-09-04 |

| Production Status | End-of-life | Active |

---

Head-to-Head Benchmarks

The direct head-to-head benchmark list is empty, so the comparison relies on each part’s individual scores. The GT 745M shows three compute results: Geekbench Metal at 2777, Geekbench OpenCL at 3580, and Geekbench Vulkan at 5502. The Vulkan result is the strongest for the Kepler GPU, indicating competitive performance in modern API workloads relative to its average.

The Arc Pro B60 has a broader suite. Its Passmark G3D score is 14580, far above the GT 745M’s highest single benchmark, the Geekbench Vulkan score of 5502. The Passmark GPU Compute score is 7029, double the GT 745M’s OpenCL score. The Arc’s 3DMark Steel Nomad DX12 score is 2646, which is a different test but shows it can handle modern DirectX 12 workloads.

The biggest win for the GT 745M is in average score. At 3953, it beats the Arc Pro B60’s 3182 by a 24.2% margin. That is despite the Arc having 2.7 times the shading units and 7 times the bandwidth. The GT 745M also leads in Geekbench Vulkan, where its 5502 exceeds any single Arc Pro B60 benchmark except the Passmark G3D and Compute results.

The Arc Pro B60 wins decisively in raw compute and graphics throughput. Its FP32 rate of 12.29 TFLOPS is 29.2 times higher than the GT 745M’s 421.6 GFLOPS. The pixel rate of 192.0 GPixel/s is 43.7 times higher than 4.392 GPixel/s. The memory bandwidth of 456.0 GB/s is 7.1 times higher than 64.00 GB/s. These are not marginal gains; they are order-of-magnitude differences.

The GT 745M’s nearest rivals average around 3955 (Radeon R5 M420, GeForce 830M), so it is well matched to its peers. The Arc Pro B60’s nearest rivals are the Quadro P1000 (3163) and GeForce GT 640 (3210), with the RTX 5080 Ti at 3075 showing a 3.5% delta. That means the Arc Pro B60 is slower than the GT 640 by 0.9% and faster than the RTX 5080 Ti by 3.5%, which is an unusual placement for a 200 W pro card.

In summary, the GT 745M is a legacy part that punches above its weight in average synthetic scores, while the Arc Pro B60 is a modern compute workhorse with vastly superior memory and compute resources. The recorded data shows no direct head-to-head benchmark, so the verdict depends on workload relevance. For DirectX 12 and Vulkan 1.4, the Arc Pro B60 is the only one of the two that supports those specifications fully. For pure OpenCL or legacy compatibility, the GT 745M still holds up statistically, but its 23rd percentile ranking suggests it is far behind most current hardware.