Release v0.23.1
tokenizers 0.23.1 is the first proper stable release in the 0.23 line — 0.23.0 only ever shipped as rc0 because the release pipeline itself was broken (Node side hadn't shipped multi-platform binaries since 2023, Python side was on pyo3 0.27 without free-threaded support). 0.23.1 is the version where everything actually goes out the door together: full Node multi-platform wheels for the first time in years, Python 3.14 (regular and free-threaded 3.14t), full type hints for every Python class, and a stack of measurable perf wins on the BPE / added-vocab hot paths.
There is no functional 0.23.0 published — we tag 0.23.1 directly so users don't accidentally pull a never-shipped version.
requires-python = ">=3.10"; 3.9 users stay on 0.22.x.add_tokens normalizes content at insertion (#1995) — re-saved tokenizer.json may differ in the added_tokens block. Existing files load unchanged.Any now return real types; mypy --strict may surface previously-hidden errors. Stub layout also moved from tokenizers/<sub>/__init__.pyi to tokenizers/<sub>.pyi. This breaks the surface of some of the processors like RobertaProcessign's __init__ .PyResult<T> because of Arc<RwLock<Tokenizer>>; a poisoned lock surfaces as PyException instead of a panic.Run with cargo bench --bench <name> -- --save-baseline v0_22_2 on v0.22.2, then --baseline v0_22_2 on v0.23.1. Numbers are point-in-time wall clock on a single laptop; relative deltas are what matters, absolute numbers will differ on CI hardware.
bench: improve added_vocab_deserialize to reflect real-world workloads (#2000) is now representative of how transformers actually loads tokenizer.json files. The combined effect of daachorse for the matching automaton plus the normalize-on-insert refactor is enormous on this workload:
| benchmark | v0.22.2 | v0.23.1 | change |
|---|---|---|---|
| 100k tokens, special, no norm | ~410 ms | 248 ms | −40% |
| 100k tokens, non-special, no norm | ~7.1 s | 273 ms | −96% |
| 100k tokens, special, NFKC | ~395 ms | 235 ms | −40% |
| 100k tokens, non-special, NFKC | ~7.4 s | 290 ms | −96% |
| 400k tokens, special, no norm | ~15 s | 980 ms | −94% |
Real-world impact: loading a Llama-3-style tokenizer with a large set of added tokens dropped from "noticeable pause" to "instant".
| benchmark | v0.22.2 | v0.23.1 | change |
|---|---|---|---|
BPE GPT2 encode batch, no cache | 530 ms | 446 ms | −16% |
BPE GPT2 encode batch (cached) | 690 ms | 685 ms | noise |
BPE GPT2 encode (single) | 1.95 s | 1.94 s | noise |
BPE Train (small) | 32.6 ms | 31.5 ms | −3% |
BPE Train (big) | 1.01 s | 988 ms | −2% |
The BPE per-thread cache PR (#2028) shows much larger wins on highly-parallel workloads (+47–62% at 88+ threads on a server box, per the PR's own measurements on Vera). Single-thread batch numbers above are flat or slightly improved because cache-hit overhead was already low without contention.
| benchmark | v0.22.2 | v0.23.1 | change |
|---|---|---|---|
llama3-encode (single) | 2.10 s | 2.02 s | −4% |
llama3-batch | 438 ms | 408 ms | −7% |
llama3-offsets | 410 ms | 395 ms | −4% |
Right-direction truncation no longer pre-tokenizes past max_length. The new truncation_benchmark doesn't exist on v0.22.2 so there's no apples-to-apples here, but the PR's own measurements on the same machine showed −20–28% across a range of max_length values for right-truncation; left-truncation unchanged.
BPE::Builder::build no longer formats strings in a hot loop (#2010) — ~45% faster Tokenizer::from_file on Llama-3 in the PR's profile.The tokenizer.json format is forward-compatible: existing files load on 0.23 unchanged. Two things to know if you re-save:
added_tokens entries created via add_tokens(..., normalized=True) will have their content normalized at save time — see breaking-change note above.tokenizer.train(...) no longer keeps a redundant added_tokens/special_tokens Vec separate from the added_tokens_map_r. Public API surface unchanged; only the internal struct shape moved.bench: improve added_vocab_deserialize to reflect real-world workloads (#2000) lands a more realistic micro-benchmark for this surface; if you're tracking deserialize perf in your own CI, the new bench is the one to compare against.
Dedicated wheels for python3.14t (the free-threaded build introduced in PEP 703). The wheel:
Py_MOD_GIL_NOT_USED, so importing tokenizers does not force the GIL back on.abi3 cargo feature (free-threaded Python doesn't expose the limited API).Arc<RwLock<Tokenizer>> for the inner state so concurrent setters and encoders don't race PyO3's per-pyclass borrow check.A new stress-test module tests/test_freethreaded.py exercises N-encoder × M-setter races on a single Tokenizer and asserts no RuntimeError: Already borrowed, no RwLock poisoning, and that sys._is_gil_enabled() is False post-import.
For the regular CPython wheel everything is unchanged.
The npm package now ships 13 platforms (macOS x64/arm64/universal, Windows x64/i686/arm64, Linux x64/arm64/armv7 in both glibc and musl, Android arm64/armv7) — previous workflows only built 3 of those, leaving Apple Silicon / Linux ARM / Alpine users with package-not-found errors since 2023 (#1365, #1703, #1922). Fixed via #1970 + #2034, which also bumps @napi-rs/cli to v3 and switches cross-builds to cargo-zigbuild.
Every class in the python bindings now ships proper .pyi stubs — Tokenizer, AddedToken, Encoding, every decoder / model / normalizer / pre-tokenizer / processor / trainer. Editors and type checkers (mypy, pyright, ty) see real signatures with types and docstrings instead of falling back to Any.
The stubs are generated automatically from the compiled extension via tools/stub-gen (Rust binary using pyo3-introspection). Re-running make style regenerates them; CI guards against regenerated-vs-checked-in drift. If the generator ever returns 0 docstrings (e.g. because the [patch.crates-io] pin in .cargo/config.toml falls out of sync with the pyo3 dep version), it now hard-aborts with a precise diagnostic instead of silently emitting bare-bones stubs.
>>> from tokenizers import Tokenizer
>>> # IDEs now resolve every method, every kwarg, every return type
>>> Tokenizer.from_pretrained("bert-base-cased")⚠️ As called out in breaking changes: stricter type info means previously-hidden type errors in user code may now surface under mypy --strict.
models.Unigram now exposes alpha and nbest_size for subword regularization (parity with Google's implementation, #1994). Closes long-standing requests #730 and #849.Tokenizer (#1958) — useful for long-lived caches that don't want to keep tokenizers alive.ci_benchmark against the stored baseline and posts a comparison chart to the PR.EncodingVisualizer: unclosed annotation span fixed (#1911), HTML escape applied to output (#1937).__copy__ / __deepcopy__ (#1930).to_vec() from slice (#1964).wget / norvig URL with HF Hub downloads in test data fetch (#2018).uv support in the Python Makefile (#1977).Thanks to everyone who shipped commits between v0.22.2 and v0.23.1:
@ArthurZucker, @finnagin, @gordonmessmer, @jberg5, @kennethsible, @llukito, @MayCXC, @McPatate, @michaelfeil, @mrkm4ntr, @musicinmybrain, @ngoldbaum, @OhashiReon, @paulinebm, @podarok, @rtrompier, @sebpop, @Shivam-Bhardwaj, @threexc, @wheynelau, @xanderlent — plus @dependabot and @hf-security-analysis for keeping pins fresh.
Full Changelog: https://github.com/huggingface/tokenizers/compare/v0.22.2...v0.23.1
Fetched April 27, 2026