TensorRT-LLM removed its TensorRT engine backend in the 1.3 release-candidate line (from v1.3.0rc21, 2026-07-15); PyTorch is the only backend there
Removed in 1.3.0rc21 pre-release; stable 1.2.1 retains itobserved
| Value kind | observed — Reported measurements, counts, and specifications keep the precision and scope stated by their source; an exact specification is not treated as a range. |
|---|---|
| Scope | Removed from v1.3.0rc21 (2026-07-15), a pre-release; the 1.4 pre-release line (v1.4.0rc0, 2026-10-14) is also PyTorch-only. The latest stable release, v1.2.1 (2026-04-20), still has the engine-build workflow; 1.3.0 had not reached general availability as of 2026-10-19. NVIDIA's release-notes page lists the removal under 'Release 1.2'. The migration page documents the new workflow; it does not rank engines. |
| As of | 2026-10-19 |
| Source | NVIDIA TensorRT-LLM GitHub releases v1.3.0rc20 (2026-06-30) and v1.3.0rc21 (2026-07-15); Migration Guide: TensorRT Backend Removed, page last updated 2026-08-27. · v1.3.0rc21 release notes: 'Remove Python modules and tests for the legacy TensorRT backend'; stable line v1.2.1 (2026-04-20) |
| Review | checking…review by 2027-01-31 · standard cadence |
| Recorded changes | last 2026-10-19 |
| Claim id | part10-tensorrt-llm-pytorch-only-backend |
Where the guide uses it
Not quoted in a chapter yet; it is kept in the curated register.
← Full numbers register — every date-stamped figure in the guide, with revision history.