NVIDIA GeForce GTX 970 vs NVIDIA GeForce GTX 980 Comparison

Head-to-Head Benchmarks

The benchmark data leaves no ambiguity: the NVIDIA GeForce GTX 980 wins all 11 recorded head-to-head tests against the GTX 970. The largest margin comes in the 3DMark Steel Nomad DX12 test, where the GTX 980 scores 474 against the GTX 970’s 369, a 28.5% advantage. This is the only test in the entire suite where the delta exceeds 20%, and it signals that the GTX 980 handles modern DirectX 12 workloads with considerably more headroom.

In compute-oriented benchmarks, the GTX 980 maintains a consistent lead. Geekbench OpenCL shows a 15.7% gap (34676 vs 29982), while PassMark GPU Compute posts a 16.7% difference (4753 vs 4073). These results align with the raw FP32 throughput figures: the GTX 980 delivers 4.981 TFLOPS versus 3.920 TFLOPS for the GTX 970, a difference that manifests directly in general-purpose GPU workloads.

The synthetic graphics tests tell a similar story. PassMark G3D shows the GTX 980 at 11095 versus 9638, a 15.1% edge. The DirectX 11 test narrows slightly to 16.9% (83 vs 71), while DirectX 10 comes in at 15.2% (53 vs 46). Even legacy DirectX 9 performance favors the GTX 980 by 14.7% (164 vs 143). The smallest margins appear in 2D performance, where PassMark G2D shows only a 3.7% difference (792 vs 764), suggesting that both cards are largely limited by the same memory subsystem in 2D workloads.

Vulkan and Metal results also favor the GTX 980, with 12.7% (22543 vs 20005) and 11.5% (15163 vs 13595) gaps respectively. The DirectX 12 result from PassMark shows a 12.2% difference (46 vs 41), reinforcing the pattern that the GTX 980’s advantage scales with compute intensity rather than API generation. Across all 11 tests, the GTX 980 wins every encounter, with an average benchmark score of 8167 against 7157 for the GTX 970.

It is worth noting how the two cards sit relative to their nearest rivals in the database. The GTX 980’s average score places it just 0.4% above the NVIDIA GeForce GTX 950M (8135) and 0.5% above the AMD Radeon R9 M360 (8129), which means its 43rd percentile ranking reflects a crowded mid-range field. The GTX 970, by contrast, sits 0.2% below the Intel Iris Pro Graphics P580 (7170) and 0.9% below the NVIDIA GeForce GTX 750 (7222), placing it at the 39th percentile. The head-to-head margin between the two cards is therefore more meaningful than their absolute percentile rankings suggest.

The Verdict

The data indicates a clear performance hierarchy: the GTX 980 outclasses the GTX 970 in every recorded benchmark. The 28.5% lead in 3DMark Steel Nomad DX12 is the standout result, but even the smallest margin (3.7% in PassMark G2D) favors the GTX 980. For users prioritizing maximum frame rates in compute-heavy or modern API workloads, the GTX 980 is the unambiguous choice based on the recorded scores.

That said, the GTX 970 is not without rationale in the database. Its average benchmark score of 7157 sits within 0.9% of the NVIDIA GeForce GTX 750 (7222), meaning it performs comparably to that older card despite being a newer design. The GTX 970 also carries a lower thermal envelope at 148 W TDP versus 165 W for the GTX 980, and its suggested PSU requirement is 300 W versus 450 W. For systems with limited power headroom, the GTX 970 represents the more restrained option, though it sacrifices between 11.5% and 28.5% of performance depending on the workload.

The GTX 980’s launch MSRP is 549 USD, while the GTX 970 launched at 329 USD. Neither card is in current production, and both are marked as end-of-life in the database. The performance gap is consistent and substantial, but the power and PSU differences may matter for builds where thermal or electrical constraints are tight.

Architecture Differences

Both cards share the same fundamental architecture: Maxwell 2.0 on the GM204 chip, fabricated on TSMC’s 28 nm process with 5,200 million transistors on a 398 mm² die. The transistor density is identical at 13.1M per mm². The memory configuration is also the same: 4 GB of GDDR5 on a 256-bit bus with 224.4 GB/s of bandwidth, running at 7 Gbps effective.

The divergence lies in the execution resources. The GTX 980 packs 2048 shading units, 128 texture mapping units, and 64 ROPs. The GTX 970 reduces these to 1664 shading units, 104 TMUs, and 56 ROPs. This translates to a 384-shader deficit for the GTX 970, which is the primary reason for its lower throughput. The pixel rate reflects this: 77.82 GPixel/s for the GTX 980 versus 65.97 GPixel/s for the GTX 970, a 17.9% gap. Texture rate shows a similar trend at 155.6 GTexel/s versus 122.5 GTexel/s, an 27% difference.

Clock speeds also differ. The GTX 980 runs a base clock of 1127 MHz with a boost of 1216 MHz, while the GTX 970 operates at 1050 MHz base and 1178 MHz boost. The GTX 980’s higher clocks compound its shader advantage, producing the 4.981 TFLOPS FP32 figure versus 3.920 TFLOPS. Both cards use the same dual-slot form factor, measure 267 mm in length, 111 mm in height, and 40 mm in width, and require two 6-pin power connectors.

Feature support is identical across both cards: DirectX 12 (12_1), OpenGL 4.6, Vulkan 1.4, and the same display outputs (1x DVI, 1x HDMI 2.0, 3x DisplayPort 1.2). Neither card includes ray tracing cores or tensor cores, and both use the PCIe 3.0 x16 bus interface. The release dates are the same (2014-09-18), and both belong to the GeForce 900 generation, succeeding the GeForce 700 series and preceding GeForce 10.

FAQ

Q: Which card has a higher average benchmark score?

A: The GTX 980 records an average benchmark score of 8167, while the GTX 970 averages 7157, a difference of roughly 14%.

Q: What is the largest performance gap between the two cards?

A: The largest gap is in the 3DMark Steel Nomad DX12 test, where the GTX 980 leads by 28.5% (474 vs 369).

Q: Do both cards use the same memory configuration?

A: Yes, both have 4 GB of GDDR5 on a 256-bit bus with 224.4 GB/s of bandwidth and 7 Gbps effective memory speed.

Q: Which card draws less power?

A: The GTX 970 has a lower TDP of 148 W versus 165 W for the GTX 980, and its suggested PSU is 300 W compared to 450 W.

Q: Are there any benchmark tests where the GTX 970 wins?

A: No, the database records 11 head-to-head tests, and the GTX 980 wins all 11.

Q: What is the difference in shading units?

A: The GTX 980 has 2048 shading units, while the GTX 970 has 1664, a deficit of 384 units.

Where Each One Wins

The GTX 980 wins in every measured category, but the magnitude of its advantage varies by workload type. For modern DirectX 12 gaming, the 3DMark Steel Nomad result (28.5% lead) makes it the clear pick. Compute-heavy tasks such as OpenCL workloads (15.7% lead) and GPU compute benchmarks (16.7% lead) also strongly favor the GTX 980, which aligns with its higher FP32 throughput of 4.981 TFLOPS versus 3.920 TFLOPS.

The GTX 970’s only advantages are indirect: lower TDP (148 W vs 165 W), lower suggested PSU requirement (300 W vs 450 W), and a launch MSRP of 329 USD versus 549 USD. In scenarios where power delivery is constrained, such as compact builds with modest power supplies, the GTX 970 is the more feasible option despite its performance deficit. The 3.7% gap in PassMark G2D also suggests that 2D desktop workloads are nearly equivalent, so users who rarely push 3D workloads may not notice the difference.

For Vulkan and Metal users, the GTX 980 leads by 12.7% and 11.5% respectively, making it the better choice for cross-platform development or Vulkan-based titles. The DirectX 11 result (16.9% lead) and DirectX 10 result (15.2% lead) confirm that the GTX 980’s advantage is not limited to new APIs. In short, the GTX 980 is the superior performer across the board, while the GTX 970 is the power-conscious alternative that sacrifices measurable performance for efficiency.