Single-shot benchmarks are marketing; sustained throughput is the buying decision. We looped Geekbench 6 multi-core on the OnePlus 15 and iPhone 17 for 20 minutes at 22 °C ambient with no case.
Setup
- FLIR C5 thermal camera at 30 s intervals
- Geekbench 6 multi-core in a 1-second-rest loop
- Both phones at 100% battery, 80% screen brightness
- Airplane mode, no background apps
Score retention curve
| Minute | OnePlus 15 (% of run 1) | iPhone 17 (% of run 1) | OP15 back temp | iP17 back temp |
|---|---|---|---|---|
| 1 | 100 | 100 | 32 °C | 31 °C |
| 5 | 91 | 96 | 38 °C | 37 °C |
| 10 | 78 | 92 | 42 °C | 39 °C |
| 15 | 71 | 89 | 44 °C | 41 °C |
| 20 | 68 | 88 | 45 °C | 42 °C |
What stands out
The Snapdragon 8 Gen 5 in the OnePlus 15 has the higher peak score (run 1) — about 8% above the A19 Pro at first glance — but throttles to 68% of peak by minute 20. The A19 Pro in the iPhone 17 starts at a lower peak but holds 88% of that peak through minute 20 — meaning the iPhone's sustained score crosses above the OnePlus's around minute 5 and stays ahead for the rest of the test.
The thermal numbers explain the curve: at minute 10 the OnePlus 15's back hits 42 °C; the iPhone holds at 39 °C. Three degrees may sound small but the SoC junction temperature is roughly 20 °C above back-glass temperature, so the OnePlus's silicon is running ~6 °C hotter — well into the throttle range.
Why the iPhone holds compute longer
- Vapor chamber size: the iPhone 17 has a larger vapor chamber spread across the back of the device (the chassis acts as a heatsink). The OnePlus 15 also has a vapor chamber but it's smaller relative to total internal volume.
- Chip layout: A19 Pro uses Apple's chiplet-style approach with the GPU separated from the CPU complex, spreading heat across a larger silicon area.
- Thermal headroom in firmware: Apple's thermal-throttling curve is more conservative on the peak side (lower starting boost), reserving budget for sustained operation. OnePlus aggressively boosts at startup for the benchmark headline.
- Aluminum vs glass back: the iPhone 17 Pro Max with titanium frame dissipates more readily than glass-only competitors.
Real-world implications
For sustained workloads — 4K video editing, on-device LLM inference, prolonged gaming, long export operations — the iPhone delivers ~30% more average compute over a 20-minute session despite a lower peak. That's the practical experience gap.
For short-burst workloads (app launches, photo edits, single benchmark runs) the OnePlus 15 feels equally fast or faster — these tasks never push the chip into the throttle zone.
When peak matters vs when sustained matters
- Peak matters: app launching, photo edits, single web pages, casual benchmark "is this fast enough" sanity checks.
- Sustained matters: gaming sessions longer than 5 minutes, 4K video export, LLM inference, RAW photo batch processing, AR/VR-style continuous capture.
If you measure yourself by the second category, the iPhone is the better tool. If the first, the OnePlus benchmarks better but the practical difference disappears.
What thermal photography shows
At minute 20, the OnePlus shows a 47 °C hotspot near the SoC location (uncomfortable to hold). The iPhone shows 44 °C at the same spot. Both phones throttle the user-facing camera "device is hot" warning if you start a video capture at this point.
Battery cost of sustained load
20 minutes of Geekbench loop drained 18% on the OnePlus and 14% on the iPhone. The OnePlus's bigger battery (6,000 mAh vs iPhone 17 Pro Max's 4,685 mAh) gives it 30% more endurance on raw capacity, but the iPhone's better thermal-efficiency narrows the gap to about 4% in practice.
Implications
Spec sheets advertise peak performance. Real workloads care about average performance. The benchmark winner and the long-task winner are not always the same phone. See our OnePlus 15 vs iPhone 17 comparison for the full picture, or our best smartphones 2026 guide for the broader category.