NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI, #4828 most downloaded on PyPI
Massive Text Embedding Benchmark
Latest stable
2.27.8
Last release 12 days ago
10 Nov 2026
Ships on a steady schedule
a new release about every 8 days
Nearly every release is documented
notes for 60 of the last 60 stable releases
Nothing withdrawn
no release was ever pulled
4 years old
875 releases · first in 2022
Drag across the ruler to compare, or tap a release to open it.
add models from collection and revisions
fix: Add WebSSL models (#2604)
add 2 web SSL dino models
add models from collection and revisions
update memory_usage_mb and embed dim
use automodel instead (afb72ac)
fix mieb citation (#2606) (5a74754)
update Doubao-1.5-Embedding (#2575)
update seed-embedding
update seed models
fix linting and tiktoken problem
fix tiktoken bug
fix lint
update name
Update mteb/models/seed_models.py
adopt suggestion
Co-authored-by: Roman Solomatin <samoed.roman@gmail.com>
update logging
update lint
Co-authored-by: zhangpeitian <zhangpeitian@bytedance.com>
Co-authored-by: Roman Solomatin <samoed.roman@gmail.com> (0bda363)
Add image only MIEB benchmark to LB left panel (#2596)
Update benchmarks.py
make lint
add to left side bar (039a965)
Add MIEB image only benchmark (#2590)
add vision only bench
add description
correct zs task modalities
specify tasks param (7b6d9d7)
fix codecarbon version (#2587) (ca10bac)
fix FlagEmbedding package name (#2588) (b1606ff)
Your coding agent can read these notes before it upgrades. Set up the MCP server →
Remove the comments from ImageEncoder (#2579) (`951bae3`)
fix: Add Encodechka benchmark (#2561)
add tasks
add benchmark
fix imports
update stsb split (0737e78)
Update tasks table (4f23d62)
Remove the comments from ImageEncoder (#2579) (951bae3)
move icon & name to benchmark dataclass (#2573) (fa5f034)
Update tasks table (e03333f)
Add missing annotations (#2498) (235906b)
Docs: Improve MIEB docs (#2569) (adfd92a)
Add ModelMeta for CodeSearch-ModernBERT-Crow-Plus (#2570)
Add files via upload
Update shuu_model.py
Update shuu_model.py
Update shuu_model.py
Update shuu_model.py
Update shuu_model.py
Update shuu_model.py
Update shuu_model.py
Update shuu_model.py
Update shuu_model.py
Update shuu_model.py
Update shuu_model.py
Update shuu_model.py
Update overview.py
Update shuu_model.py
Update shuu_model.py
Update shuu_model.py
Update mteb/models/shuu_model.py
Co-authored-by: Roman Solomatin <samoed.roman@gmail.com>
Co-authored-by: Roman Solomatin <samoed.roman@gmail.com> (713635a)
Update tasks table (e56aab5)
Backfill task metadata for metadata for GermanDPR and GermanQuAD (#2566)
Add metadata for GermanDPR and GermanQuAD
PR improvements (c0d3ca0)
Add relle (#2564)
Add relle
defined model metadata for relle
Add mteb/models/relle_models.py
Update mteb/models/relle_models.py
Co-authored-by: Roman Solomatin <samoed.roman@gmail.com>
run after "make lint"
Add model into model_modules and lint check
Co-authored-by: Roman Solomatin <samoed.roman@gmail.com> (f11ac2a)
d475c7e)fix: jasper models embeddings having nan values (#2481) (`f7072d5`)
f7072d5)Fix leaderboard entry for BuiltBench (#2562)
Co-authored-by: Mehrzad Shahin-Moghadam <mehr@Mehrzads-MacBook-Pro.local> (4b755a3)
add USER2 (#2560)
add user2
add training code
update prompts (5ed6773)
Bumped gradio version to latest
feat: UI Overhaul (#2549)
Bumped gradio version to latest
Added new Gradio table functionality to leaderboard
Removed search bar
Changed color scheme in plot to match the table
Added new benchmark selector in sidebar
Changed not activated button type to secondary
Short-circuited callbacks that are based on language selection
Re-added column width calculation since it got messed up
Commented out gradient for per-task table as it slowed things down substantially
Styling and layout updates
Adjusted comments according to reviews
Converted all print statements to logger.debug
Removed pydantic version fix
Ran linting
Remove commented out code
Co-authored-by: Kenneth Enevoldsen <kennethcenevoldsen@gmail.com>
Moved English,v1 to Legacy section
Closed the benchmark sharing accordion by default
Adjusted markdown blocks according to suggestions
Ran linter
Co-authored-by: Kenneth Enevoldsen <kennethcenevoldsen@gmail.com> (0ab947b)
fix unintentional working of filters on leaderboard (#2535)
fix unintentional working of filters on leaderboard
address comments
make lint
address comments
rollback unnecessary changes (50d7e9e)
fix e5_R_mistral_7b (#2490)
fix e5_R_mistral_7b
change wrapper
address comments
Added kwargs for pad_token
correct lang format
address comments
add revision
Co-authored-by: Roman Solomatin <samoed.roman@gmail.com> (4a6e539)
feat: Added dataframe utilities to BenchmarkResults
feat: Added dataframe utilities to BenchmarkResults (#2542)
fix: Added dataframe utilities to BenchmarkResults
get_results_table. I was considering renaming it to to_dataframe to align with tasks.to_dataframe. WDYT?Prerequisite for #2454:
@ayush1298 can I ask you to review this PR as well? I hope this give an idea of what I was hinting at. Sorry that it took a while. I wanted to make sure to get it right.
refactor to to_dataframe and combine common dependencies
ibid
fix revision joining after discussion with @x-tabdeveloping
remove strict=True for zip() as it is a >3.9 feature
updated mock cache (8fe5742)
fix me5 trainind data config to include xquad dataset (#2552)
fix: me5 trainind data config to include xquad dataset
Update mteb/models/e5_models.py
upddate: xquad key name
Co-authored-by: Roman Solomatin <samoed.roman@gmail.com>
Co-authored-by: Roman Solomatin <samoed.roman@gmail.com> (1f82b59)
Add xlm_roberta_ua_distilled (#2547)
defined model metadata for xlm_roberta_ua_distilled
Update mteb/models/ua_sentence_models.py
Co-authored-by: Roman Solomatin <samoed.roman@gmail.com>
included ua_sentence_models.py in overview.py
applied linting, added missing fields in ModelMeta
applied linting
Co-authored-by: Roman Solomatin <samoed.roman@gmail.com> (3ff993d)
Add MIEB to README (75d3597)
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
One column per quarter.