Checklist
What happened?
libmamba reports PackagesNotFoundError for python=3.12 from defaults on win-64, while conda search sees it and --solver classic succeeds
Conda Info
active environment : None
shell level : 0
user config file : C:\Users\顾籽黎\.condarc
populated config files : C:\D\miniconda3\.condarc
C:\D\miniconda3\condarc.d\anaconda-auth.yml
conda version : 26.1.1
conda-build version : not installed
python version : 3.13.11.final.0
solver : libmamba (default)
virtual packages : __archspec=1=skylake
__conda=26.1.1=0
__cuda=12.8=0
__win=10.0.26200=0
base environment : C:\D\miniconda3 (read only)
conda av data dir : C:\D\miniconda3\etc\conda
conda av metadata url : None
channel URLs : https://repo.anaconda.com/pkgs/main/win-64
https://repo.anaconda.com/pkgs/main/noarch
https://repo.anaconda.com/pkgs/r/win-64
https://repo.anaconda.com/pkgs/r/noarch
https://repo.anaconda.com/pkgs/msys2/win-64
https://repo.anaconda.com/pkgs/msys2/noarch
package cache : C:\D\miniconda3\pkgs
C:\Users\顾籽黎\.conda\pkgs
C:\Users\顾籽黎\AppData\Local\conda\conda\pkgs
envs directories : C:\Users\顾籽黎\.conda\envs
C:\D\miniconda3\envs
C:\Users\顾籽黎\AppData\Local\conda\conda\envs
temporary directory : C:\Users\顾籽黎\AppData\Local\Temp
platform : win-64
user-agent : conda/26.1.1 requests/2.32.5 CPython/3.13.11 Windows/11 Windows/10.0.26200 solver/libmamba conda-libmamba-solver/25.11.0 libmambapy/2.3.2 aau/0.7.5 c/. s/. e/.
administrator : False
netrc file : None
offline mode : False
Conda Config
==> C:\D\miniconda3\.condarc <==
channels:
- defaults
==> C:\D\miniconda3\condarc.d\anaconda-auth.yml <==
channel_settings:
- frozendict.frozendict({'channel': 'https://repo.anaconda.cloud/*', 'auth': 'anaconda-auth'})
Conda list
# packages in environment at C:\D\miniconda3:
#
# Name Version Build Channel
anaconda-anon-usage 0.7.5 pyhb46e38b_100 defaults
anaconda-auth 0.12.3 py313haa95532_0 defaults
anaconda-cli-base 0.7.0 py313haa95532_0 defaults
anaconda_powershell_prompt 1.1.0 haa95532_1 defaults
anaconda_prompt 1.1.0 haa95532_1 defaults
annotated-types 0.6.0 py313haa95532_1 defaults
archspec 0.2.5 pyhd3eb1b0_0 defaults
boltons 25.0.0 py313haa95532_0 defaults
brotlicffi 1.2.0.0 py313h885b0b7_0 defaults
bzip2 1.0.8 h2bbff1b_6 defaults
ca-certificates 2025.12.2 haa95532_0 defaults
certifi 2026.01.04 py313haa95532_0 defaults
cffi 2.0.0 py313h02ab6af_1 defaults
charset-normalizer 3.4.4 py313haa95532_0 defaults
click 8.2.1 py313haa95532_1 defaults
colorama 0.4.6 py313haa95532_0 defaults
conda 26.1.1 py313haa95532_0 defaults
conda-anaconda-telemetry 0.3.0 pyhd3eb1b0_1 defaults
conda-anaconda-tos 0.2.2 py313haa95532_1 defaults
conda-content-trust 0.2.0 py313haa95532_1 defaults
conda-libmamba-solver 25.11.0 pyhdf14ebd_0 defaults
conda-package-handling 2.4.0 py313haa95532_1 defaults
conda-package-streaming 0.12.0 py313haa95532_1 defaults
cpp-expected 1.1.0 h214f63a_0 defaults
cryptography 46.0.3 py313habbc9f9_0 defaults
distro 1.9.0 py313haa95532_0 defaults
fmt 11.2.0 h58b7f6e_0 defaults
frozendict 2.4.6 py313h02ab6af_0 defaults
idna 3.11 py313haa95532_0 defaults
jaraco.classes 3.4.0 py313haa95532_0 defaults
jaraco.context 6.0.0 py313haa95532_0 defaults
jaraco.functools 4.1.0 py313haa95532_0 defaults
jsonpatch 1.33 py313haa95532_1 defaults
jsonpointer 3.0.0 py313haa95532_0 defaults
keyring 25.7.0 py313haa95532_0 defaults
libarchive 3.8.2 h6c023e8_0 defaults
libcurl 8.16.0 h97e0424_0 defaults
libexpat 2.7.3 h885b0b7_4 defaults
libffi 3.4.4 hd77b12b_1 defaults
libiconv 1.16 h2bbff1b_3 defaults
libmamba 2.3.2 hc213065_1 defaults
libmambapy 2.3.2 py313h364efb6_1 defaults
libmpdec 4.0.0 h827c3e9_0 defaults
libsolv 0.7.30 h23a355e_2 defaults
libssh2 1.11.1 h2addb87_0 defaults
libxml2 2.13.9 h6201b9f_0 defaults
libzlib 1.3.1 h02ab6af_0 defaults
lz4-c 1.9.4 h2bbff1b_1 defaults
markdown-it-py 4.0.0 py313haa95532_0 defaults
mdurl 0.1.2 py313haa95532_0 defaults
menuinst 2.4.2 py313h885b0b7_1 defaults
more-itertools 10.8.0 py313haa95532_0 defaults
msgpack-python 1.1.1 py313h5da7b33_0 defaults
nlohmann_json 3.11.2 h6c2663c_0 defaults
openssl 3.5.5 hbb43b14_0 defaults
packaging 25.0 py313haa95532_1 defaults
pcre2 10.46 h5740b90_0 defaults
pip 25.3 pyhc872135_0 defaults
pkce 1.0.3 py313haa95532_0 defaults
platformdirs 4.5.0 py313haa95532_0 defaults
pluggy 1.5.0 py313haa95532_0 defaults
pybind11-abi 5 hd3eb1b0_0 defaults
pycosat 0.6.6 py313h827c3e9_2 defaults
pycparser 2.23 py313haa95532_0 defaults
pydantic 2.12.4 py313haa95532_0 defaults
pydantic-core 2.41.5 py313h114bc41_1 defaults
pydantic-settings 2.12.0 py313haa95532_0 defaults
pygments 2.19.2 py313haa95532_0 defaults
pyjwt 2.10.1 py313haa95532_1 defaults
pysocks 1.7.1 py313haa95532_1 defaults
python 3.13.11 h260b955_100_cp313 defaults
python-dotenv 1.1.0 py313haa95532_0 defaults
python_abi 3.13 3_cp313 defaults
pywin32-ctypes 0.2.2 py313haa95532_0 defaults
readchar 4.2.1 py313haa95532_0 defaults
reproc 14.2.4 hd77b12b_2 defaults
reproc-cpp 14.2.4 hd77b12b_2 defaults
requests 2.32.5 py313haa95532_1 defaults
rich 14.2.0 py313haa95532_0 defaults
ruamel.yaml 0.18.16 py313hb9a58be_0 defaults
ruamel.yaml.clib 0.2.14 py313hb9a58be_0 defaults
semver 3.0.4 py313haa95532_0 defaults
setuptools 80.9.0 py313haa95532_0 defaults
shellingham 1.5.4 py313haa95532_0 defaults
simdjson 3.10.1 h214f63a_0 defaults
sqlite 3.51.0 hda9a48d_0 defaults
tk 8.6.15 hf199647_0 defaults
tomli 2.2.1 py313haa95532_0 defaults
tqdm 4.67.1 py313h4442805_1 defaults
truststore 0.10.1 py313haa95532_1 defaults
typer 0.20.0 py313haa95532_0 defaults
typer-slim 0.20.0 py313haa95532_0 defaults
typer-slim-standard 0.20.0 py313haa95532_0 defaults
typing-extensions 4.15.0 py313haa95532_0 defaults
typing-inspection 0.4.2 py313haa95532_0 defaults
typing_extensions 4.15.0 py313haa95532_0 defaults
tzdata 2025b h04d1e81_0 defaults
ucrt 10.0.22621.0 haa95532_0 defaults
urllib3 2.6.1 py313haa95532_0 defaults
vc 14.3 h2df5915_10 defaults
vc14_runtime 14.44.35208 h4927774_10 defaults
vs2015_runtime 14.44.35208 ha6b5a95_10 defaults
wheel 0.45.1 py313haa95532_0 defaults
win_inet_pton 1.1.0 py313haa95532_1 defaults
xz 5.6.4 h4754444_1 defaults
yaml-cpp 0.8.0 hd77b12b_1 defaults
zlib 1.3.1 h02ab6af_0 defaults
zstandard 0.24.0 py313he335c29_0 defaults
zstd 1.5.7 h56299aa_0 defaults
Additional Context
Body
Checklist
Summary
On Windows 11 (win-64), conda create -n test python=3.12 -c defaults -y fails with:
PackagesNotFoundError: The following packages are not available from current channels:
- python=3.12
However:
conda search -c defaults python=3.12 clearly shows multiple available builds from pkgs/main
conda search -c defaults python=3.12 --info shows valid package metadata and URLs
conda create -n test python=3.12 -c defaults --solver classic -y succeeds
- My
base environment itself is installed successfully and is functional (currently running Python from base without issue)
Because of this, it looks like a conda-libmamba-solver / libmamba solving bug or incorrect error classification, rather than an actual missing package in defaults.
Environment
OS: Windows 11
Platform: win-64
conda: 26.1.1
conda-libmamba-solver: 25.11.0
libmambapy: 2.3.2
solver: libmamba (default)
channel: defaults
Relevant conda info values:
platform : win-64
solver : libmamba
base environment : C:\D\miniconda3
channel URLs :
https://repo.anaconda.com/pkgs/main/win-64
https://repo.anaconda.com/pkgs/main/noarch
https://repo.anaconda.com/pkgs/r/win-64
https://repo.anaconda.com/pkgs/r/noarch
https://repo.anaconda.com/pkgs/msys2/win-64
https://repo.anaconda.com/pkgs/msys2/noarch
Minimal reproduction
1. Confirm that the package exists in defaults
conda search -c defaults python=3.12
This returns visible results such as:
python 3.12.0 h1d929f7_0 pkgs/main
...
python 3.12.12 h716150d_1 pkgs/main
2. Try creating an environment with the default solver (libmamba)
conda create -n test python=3.12 -c defaults -y
This fails with:
PackagesNotFoundError: The following packages are not available from current channels:
- python=3.12
3. Try the same spec with the classic solver
conda create -n test python=3.12 -c defaults --solver classic -y
This succeeds.
Additional observations
I also confirmed the package metadata is available:
conda search -c defaults python=3.12 --info
This shows valid package metadata for multiple Windows builds, including package URLs under:
https://repo.anaconda.com/pkgs/main/win-64/
For example, python-3.12.12-h716150d_1.conda is visible.
I also tried:
conda create -n test python=3.12.12 -c defaults -y
and it still failed under libmamba.
I also tried forcing full repodata:
conda create -n test python=3.12 -c defaults --repodata-fn repodata.json -y
and it still failed under libmamba.
I also tried restricting to pkgs/main only:
conda create -n test python=3.12 -c https://repo.anaconda.com/pkgs/main --override-channels -y
and it still failed under libmamba.
So this does not appear to be caused by:
- missing package metadata
- missing package on
defaults
- old conda version
- extra channels interfering
What I tried during troubleshooting
- Verified
conda search -c defaults python=3.12 returns available builds
- Verified
conda search -c defaults python=3.12 --info returns valid package metadata
- Cleared index cache:
- Retried create with default solver → still failed
- Retried with exact version
python=3.12.12 → still failed
- Retried with
--repodata-fn repodata.json → still failed
- Retried using only
https://repo.anaconda.com/pkgs/main + --override-channels → still failed
- Updated conda from
25.11.1 to 26.1.1
- Retried after update → still failed
- Retried with
--solver classic → succeeded
Expected behavior
conda create -n test python=3.12 -c defaults -y should succeed, since:
- the package is visible in
conda search
- the package metadata is valid
- the same spec succeeds with
--solver classic
Actual behavior
With the default libmamba solver, conda reports:
PackagesNotFoundError: The following packages are not available from current channels:
- python=3.12
even though the package clearly exists in defaults.
Suspected issue
This looks like one of the following:
- a
libmamba solve bug on Windows for this spec/channel combination
- incorrect handling of package metadata / constraints for
python=3.12 on win-64
- a misclassified solver failure being surfaced as
PackagesNotFoundError, even though the package actually exists
Because classic succeeds and search sees the package, this seems unlikely to be a real package availability problem.
Relevant verbose behavior
With verbose logs, libmamba appears to read repodata successfully and receives the requested spec:
The solver will handle these requests:
{
"Install": [
"python=3.12"
]
}
But it still ends with PackagesNotFoundError.
This is why I suspect the package-not-found message may be masking a deeper solver issue.
Workaround
Using the classic solver works:
conda create -n test python=3.12 -c defaults --solver classic -y
Notes
My base environment itself is installed and working correctly, and Python in base is usable.
So this does not appear to be a general inability to install/use Python from the same conda installation.
Checklist
What happened?
libmamba reports
PackagesNotFoundErrorforpython=3.12fromdefaultson win-64, whileconda searchsees it and--solver classicsucceedsConda Info
Conda Config
Conda list
Additional Context
Body
Checklist
defaultschannelconda searchcan see the package--solver classicsucceeds with the same specSummary
On Windows 11 (
win-64),conda create -n test python=3.12 -c defaults -yfails with:However:
conda search -c defaults python=3.12clearly shows multiple available builds frompkgs/mainconda search -c defaults python=3.12 --infoshows valid package metadata and URLsconda create -n test python=3.12 -c defaults --solver classic -ysucceedsbaseenvironment itself is installed successfully and is functional (currently running Python from base without issue)Because of this, it looks like a
conda-libmamba-solver/libmambasolving bug or incorrect error classification, rather than an actual missing package indefaults.Environment
Relevant
conda infovalues:Minimal reproduction
1. Confirm that the package exists in
defaultsThis returns visible results such as:
2. Try creating an environment with the default solver (
libmamba)conda create -n test python=3.12 -c defaults -yThis fails with:
3. Try the same spec with the classic solver
conda create -n test python=3.12 -c defaults --solver classic -yThis succeeds.
Additional observations
I also confirmed the package metadata is available:
This shows valid package metadata for multiple Windows builds, including package URLs under:
For example,
python-3.12.12-h716150d_1.condais visible.I also tried:
conda create -n test python=3.12.12 -c defaults -yand it still failed under
libmamba.I also tried forcing full repodata:
conda create -n test python=3.12 -c defaults --repodata-fn repodata.json -yand it still failed under
libmamba.I also tried restricting to
pkgs/mainonly:conda create -n test python=3.12 -c https://repo.anaconda.com/pkgs/main --override-channels -yand it still failed under
libmamba.So this does not appear to be caused by:
defaultsWhat I tried during troubleshooting
conda search -c defaults python=3.12returns available buildsconda search -c defaults python=3.12 --inforeturns valid package metadatapython=3.12.12→ still failed--repodata-fn repodata.json→ still failedhttps://repo.anaconda.com/pkgs/main+--override-channels→ still failed25.11.1to26.1.1--solver classic→ succeededExpected behavior
conda create -n test python=3.12 -c defaults -yshould succeed, since:conda search--solver classicActual behavior
With the default
libmambasolver, conda reports:even though the package clearly exists in
defaults.Suspected issue
This looks like one of the following:
libmambasolve bug on Windows for this spec/channel combinationpython=3.12onwin-64PackagesNotFoundError, even though the package actually existsBecause
classicsucceeds andsearchsees the package, this seems unlikely to be a real package availability problem.Relevant verbose behavior
With verbose logs,
libmambaappears to read repodata successfully and receives the requested spec:But it still ends with
PackagesNotFoundError.This is why I suspect the package-not-found message may be masking a deeper solver issue.
Workaround
Using the classic solver works:
conda create -n test python=3.12 -c defaults --solver classic -yNotes
My
baseenvironment itself is installed and working correctly, and Python inbaseis usable.So this does not appear to be a general inability to install/use Python from the same conda installation.