Apple’s transition to custom M-series chips has brought significant improvements in integrated GPU performance. This blog post compares GPU performance across Apple Silicon generations—from the original M1 through M1 Pro/Max/Ultra, M2 Pro/Max/Ultra, up to the latest M3 and M4 series—using Geekbench 6 Metal benchmark scores. We use the base M1 as a 100% performance baseline to quantify generational gains. Scores are averaged across devices to neutralize thermal differences between MacBook Airs and MacBook Pros.
A Quick Look at Apple’s M-Series Chip Graphics: M1 to M4
Apple’s M-series chips have revolutionized its Mac and iPad lineup, and a big part of that success story is their powerful integrated graphics (GPUs). Each generation has brought significant improvements, making these devices more capable for everything from everyday tasks to demanding creative work and gaming. If you’re curious about how these GPUs have evolved since 2020, here’s a quick rundown.
M1: The Game Changer
The M1 chip, launched in 2020, was Apple’s first custom-designed Mac chip and it made a huge splash. Its integrated GPU, with either 7 or 8 cores, offered a massive performance jump compared to the Intel integrated graphics previously used in Macs. A key feature was the “Unified Memory Architecture,” allowing the GPU to share memory directly with the CPU for faster and more efficient graphics processing. This meant smoother visuals, quicker photo and video editing, and even some decent gaming capabilities, all while being incredibly power-efficient.
M2: Building on Success
Released in 2022, the M2 chip was a solid step up from the M1. Its GPU could have up to 10 cores, offering an incremental performance boost and even better efficiency. The base M2 also benefited from faster memory, giving its graphics an extra kick. While not a radical redesign, the M2 GPU refined the M1’s strengths, providing more power for creative apps and improving the overall visual experience, continuing Apple’s focus on impressive performance per watt.
M3: A New Architecture for Graphics
The M3 family, arriving in late 2023, introduced a brand-new GPU architecture, marking a major leap. With up to 10 cores in the base M3, this new GPU brought features like “Dynamic Caching” for smarter memory use, plus hardware support for “ray tracing” and “mesh shading” – technologies that make games and 3D graphics look much more realistic. This was a clear signal Apple was getting even more serious about high-end graphics and gaming performance. Some have said the M3 Pro was a missed opportunity but you can decide that.
M4: Refining the Future
The M4 chip, first seen in early 2024, builds directly on the advanced graphics foundation laid by the M3. Featuring a standard 10-core GPU in its base configuration, the M4 enhances the M3’s capabilities with even faster hardware-accelerated ray tracing and adds support for AV1 video decoding for more efficient video playback. This generation continues to push performance and efficiency, making Apple’s latest devices even more capable for demanding visual tasks and setting the stage for future AI-driven experiences.
Want to Dive Deeper?
This was just a quick overview! Each of these M-series chips also has more powerful “Pro,” “Max,” and even “Ultra” versions with significantly more GPU cores and capabilities designed for the most demanding professional workloads. For a full technical breakdown of core counts, specific architectural changes, performance metrics, and how Apple uses techniques like “binning” and “UltraFusion” across these generations, stay tuned for a in-depth per variant analysis coming soon.
GPU scores below are all from Geekbench 6 GPU Metal
GPU Metal Performance Across SOC Generations
Performance Comparison Table
| Apple SoC | GPU Cores | Max Memory Bandwidth | Geekbench 6 Metal | Relative to M1 |
|---|---|---|---|---|
| M1 | 7–8 | ~68 GB/s | 31,245 | 100% |
| M1 Pro | 14–16 | 200 GB/s | 66,079 | 211% |
| M1 Max | 24–32 | 400 GB/s | 111,773 | 358% |
| M1 Ultra | 48–64 | 800 GB/s | 161,628 | 517% |
| M2 | 8–10 | 100 GB/s | 43,517 | 139% |
| M2 Pro | 16–19 | 200 GB/s | 77,783 | 249% |
| M2 Max | 30–38 | 400 GB/s | 132,931 | 425% |
| M2 Ultra | 60–76 | 800 GB/s | 209,316 | 670% |
| M3 | 10 | 100 GB/s | 46,329 | 148% |
| M3 Pro | 14–18 | 150 GB/s | 74,223 | 238% |
| M3 Max | 30–40 | 300–400 GB/s | 143,702 | 460% |
| M3 Ultra | 60–80 | 819 GB/s | 259,668 | 831% |
| M4 | 10 | 120 GB/s | 58,082 | 186% |
| M4 Pro | 20 | 273 GB/s | 111,000 | 355% |
| M4 Max | 40 | 546 GB/s | 192,532 | 616% |
Conclusion
Apple Silicon GPUs have shown substantial improvements from M1 to M4, both in performance and feature sets. The M3 Ultra, with its 80-core GPU and 819 GB/s memory bandwidth, currently offers the highest graphics performance among Apple chips, achieving a Geekbench 6 Metal score of 259,668. As applications continue to leverage these advancements, users can expect enhanced performance across various tasks.
Sources
- Geekbench 6 Metal Benchmarks Browser – Official benchmark results for Apple M1 to M4 series chips
- Apple Newsroom – Official chip announcements, including M1, M2, M3, and M4 series
- Apple MacBook Pro Product Page – Specs for M1/M2/M3/M4 Pro and Max chips
- Apple Mac Studio Product Page – Specs and architecture details for M1 Ultra, M2 Ultra, and M3 Ultra
- AnandTech: Apple Launches M3/M3 Pro/M3 Max Macs – Deep technical breakdown of GPU improvements in the M3 generation
- MacRumors Apple Silicon Guide – Timeline and comparative summaries of all Apple Silicon chips
- TechRadar: M3 Ultra Overview – M3 Ultra release coverage and architecture insights
- NotebookCheck Reviews – Independent MacBook/Mac Studio testing with Metal benchmark results
