Image: Zed Industries / zed.dev Zed 1.17.2 Goes Stable: CSV Tables, DeepSeek V4 Flash, and Lower Memory for Large Files
Zed's 1.17.2 stable release ships tabular CSV/TSV previews, DeepSeek V4 Flash and V4 Pro support with low reasoning effort, a memory reduction for large files, and improvements to Git blame, stashing, and Markdown formatting.
Zed 1.17.2 went stable on August 26, graduating the tabular data preview feature that shipped in the 1.17.0 preview earlier this month. The release also adds support for two new DeepSeek models, reduces memory usage when working with large files, and tightens up Git blame and Markdown table behavior.
Tabular previews for CSV, TSV, PSV, and SSV
Data files now open in a spreadsheet-style view rather than as raw text. Columns are sortable and resizable, you can filter rows by value, and right-clicking a cell gives you a copy option. The feature works across all four common delimiter formats: comma, tab, pipe, and semicolon separated.
For developers who work with fixture data, configuration exports, or any workflow that involves looking at structured data files in the editor, this removes a common reason to open a separate tool.
DeepSeek V4 Flash and V4 Pro
Zed now supports DeepSeek V4 Flash and V4 Pro as AI model options, both with low reasoning effort configurations. The low reasoning effort setting lets you trade some reasoning depth for faster, cheaper responses on tasks that don’t need extended chain-of-thought. This sits alongside the existing Gemini 3.7 Flash support added in this release.
Memory reduction for large files
Peak memory usage when opening large files drops by approximately the size of the file itself. The previous implementation was holding a duplicate copy of the file content in memory during certain operations; the new path eliminates it. On files above a few hundred megabytes, this is a meaningful reduction in editor memory footprint.
Git blame history at specific revisions
Git blame in Zed now supports viewing blame state at a specific revision, not just the current HEAD. You can inspect who last touched a line as of any commit in the file’s history. Two new stashing options are also available: stash only tracked changes (leaving untracked files alone) or stash only staged changes (leaving the rest of the working tree untouched).
JetBrains keymap improvements
The JetBrains keymap now includes CamelHump-style subword navigation for alt-left, alt-right, shift-alt-left, and shift-alt-right. This matches how IntelliJ-family editors move the cursor through camelCase and snake_case identifiers, stopping at each word boundary rather than skipping to the next whitespace. For developers switching between JetBrains IDEs and Zed, this closes a frequently-cited muscle-memory gap.
Markdown table formatting
Column sizing in Markdown tables is now based on actual cell content width. Previously the column widths could be inconsistent, especially in tables with short headers and long cell values. The new behavior keeps tables readable without manual alignment.
Other fixes
Several smaller issues are also addressed: SSH security keys now dismiss authentication prompts automatically when auth succeeds, indentation and cursor placement edge cases in various language modes are corrected, and language server completion filtering is improved.
Release notes for 1.17.2 are on the Zed releases page.