Steam Deck FPS Charts Explained

Valve’s new Steam Deck FPS charts help you judge whether a game is likely to meet your performance expectations. Start with the average, inspect the distribution, and then check a documented settings profile before expecting the same result.

Where to find the charts

On a game’s Steam Store page, select Steam Deck Compatibility → Learn more, then expand the FPS details.

The charts are also in the game’s Hardware Compatibility dialog in the Steam client. Valve’s November 11, 2026 stable update made the feature available on desktop and Deck.

Data can appear for Verified, Playable and Unsupported games once enough samples exist. An absent chart is not a verdict that the game runs badly.

The Verified badge remains a compatibility assessment. Valve’s current developer criteria specify a default configuration delivering 30 FPS at 800p on Deck; the badge does not promise 60 FPS.

Read the average and distribution together

Valve builds averages from anonymized, opted-in gameplay data. Histogram bars show the percentage of samples at each framerate, with multiple samples per session. They do not represent percentages of players.

For readers, a narrow cluster suggests sampled performance often falls in that range. A wide spread or low-FPS tail is a reason to investigate demanding scenes. A histogram loses the order of events: it cannot tell you whether low samples came from repeated brief dips, one long slowdown or different play sessions.

An average is not a guaranteed minimum, and the histogram is not a proper 1% low measurement. For stutter, inspect frame-time behavior or gameplay footage as well: a high average can hide individual slow frames.

Valve says it filters idle and invalid samples and includes Decks using the default system frame limit: 60 FPS on LCD and 90 FPS on OLED. Custom in-game settings are not accounted for. That default limiter does not establish a shared graphics preset.

In-game caps and VSync can shape the bars; Valve describes an OLED 90/45 FPS pattern caused by synchronization. Its consumer FAQ does not disclose the collection window, exact sample threshold or treatment of generated frames.

The settings that still matter

Resolution and render scale. Record both the output resolution and how the game renders the image. “1280×800” output can be reconstructed from a smaller internal image. Native 800p and 800p with FSR Balanced are different workloads.

Graphics settings. Record the preset and any overrides. Lighting, shadows, effects and scene detail can change the work required to draw each frame, even when resolution stays the same. A custom profile needs its individual settings beside the FPS result.

Upscaling. Identify the technique, version when known, and mode. FSR 2’s Quality, Balanced and Performance modes use different internal scaling ratios. A higher-performance mode can reduce rendering work while changing image detail; inspect the moving image before accepting that tradeoff.

Frame generation. Label it separately. Interpolated frames can raise displayed FPS without increasing the underlying game’s update rate by the same amount. Compare base FPS with base FPS, and identify totals that include generated frames.

Keep results separate from targets and caps

Label	What it means	How to read it
Measured or reported FPS	Performance observed during a test or gameplay.	Preserve the source’s average, range, dips and scene context. A range midpoint is not a measured average.
FPS target	The intended performance goal.	A recommendation or dynamic feature may aim for it without consistently reaching it.
FPS cap	A configured upper limit.	A limiter can stop FPS rising above the cap; it cannot make the game reach it.

Two examples from published Steam Deck reports

These examples were checked against the creators’ original written reports. SteamGamer did not perform new tests for this article. They describe the cited configurations and dates; they are not a guarantee for every later patch and not proof that those setups contributed to Valve’s charts.

Helldivers 2

In its April 28, 2026 report, Gaming On Steam used a Steam Deck OLED, 1280×800 output and Quality render scale. The Steam Deck preset was modified with depth of field off and anti-aliasing on. VSync and the in-game framerate limiter were off.

The author reported roughly 50 FPS in quieter moments, falling to around 30 FPS and sometimes lower in intense fights on harder difficulties. That is a scene-dependent gameplay account, not a measured 40 FPS average or a guaranteed 30 FPS floor. The output resolution also needs the render-scale setting beside it.

See Helldivers 2 Steam Deck settings on SteamGamer.

Path of Exile 2

Gaming On Steam’s September 26, 2026 report used 800p output, FSR Balanced, Medium textures, minimum shadows and High water detail, with dynamic culling targeting 60 FPS. The Deck model was not specified.

The author reported around 60 FPS during much of play, dips into the 50s in busy scenes, and the 30s and 40s during spell-heavy ritual encounters. The dynamic-culling target describes what the configuration aims for. It is neither a measured average nor a frame cap.

See Path of Exile 2 Steam Deck settings on SteamGamer.

How to use the charts when choosing settings

Set your goal. Decide whether you care most about responsiveness, steadier pacing, image quality or battery life. Use Valve’s average and distribution to decide which games warrant a closer look.
Open a complete settings report. Match the Deck model when stated, output resolution, graphics options, upscaling and limiter settings. Check the report’s date and which areas or difficulty levels were covered.
Keep each setup together. SteamGamer’s summary fields can describe different observations. Use one individual report’s settings together instead of assuming the page’s separate medians form a single tested configuration.
Check a demanding scene on your Deck. Watch FPS and frame-time behavior, change one setting at a time, and inspect image quality in motion. Choose a cap that the observed performance can sustain; setting a higher cap does not create additional performance.