# PyMOL vs ChimeraX vs Chimera vs VMD

> ChimeraX is free for academics but not for companies, PyMOL is open source but sold as a licensed build, VMD is free for research and internal business use, and UCSF Chimera shipped a release in June 2026 while calling itself legacy software. We compare all four on licence, platforms, six everyday commands, trajectories, AlphaFold models and scripting.

## Which viewer should you learn?

For most academic users starting today, the choice is between PyMOL and ChimeraX, and the deciding facts are the licence and what your lab already uses. VMD is the one to learn when your main file is a molecular dynamics trajectory rather than a single structure. UCSF Chimera is worth installing only when an old protocol depends on it. Companies should read the licence section first: each program draws the commercial line in a different place.

The four look interchangeable until you ask them to do the same thing. Superimposing one HIV-1 protease structure on another is a single command in PyMOL, ChimeraX and Chimera, and each aligns the sequences before fitting. In VMD, `measure fit` returns a transformation matrix for two atom selections you have matched yourself, and `measure rmsd` requires both selections to hold the same number of atoms.

The names cause the rest of the confusion. Chimera and ChimeraX come from the same UCSF group and still write the same three residues differently: `:25-27.A` in Chimera, `/A:25-27` in ChimeraX.

The old program has not quite gone away either. Chimera shipped a release on June 25, 2026, "fixing problems on Mac OS Tahoe", and the same site calls it "legacy software that is no longer being developed or supported". Both statements are true, which is why searches and old tutorials mix the two up.

Below, the four programs are compared on cost, platforms, six everyday tasks written in each command language, trajectories, AlphaFold models and scripting. The [PDB viewer](/app/pdb-viewer) on ProteinIQ reads PyMOL commands and a subset of ChimeraX commands in the browser, so you can try both syntaxes on a structure before installing either.

## Licences and cost

All four are free for some people and not for others, and the rules differ. ChimeraX can be downloaded free of charge for academic, government, nonprofit and personal use, and commercial users need a separately executed written licence.

The ChimeraX licence defines non-commercial research narrowly: research counts only if it is neither conducted nor funded by an entity engaged in the commercial use or exploitation of works similar to the software. A university lab with an industry sponsor should read that clause before assuming it is covered. We found no commercial price on the pages we could read, so a company has to ask UCSF.

PyMOL is a commercial product whose source is mostly open. Schrödinger says it makes "most of its source code freely available under a permissive license", and the open-source repository is published under a BSD-like licence. The installers on pymol.org are the licensed build, priced per year, with a free educational option for full-time students and educators teaching full-time students. The [PyMOL licence and price guide](/guides/pymol-license) goes through each tier.

VMD sits between the two. The University of Illinois distributes it free of charge with source code, under a licence that grants use "for academic, research and internal business purposes only", requires a commercial licence for commercial use and limits redistribution. Downloading requires registration, and published work that uses VMD is expected to cite Humphrey, Dalke and Schulten's 1996 paper. Table 1 sets the four side by side.

  <caption><strong>Table 1. Licence, cost and platforms of PyMOL, ChimeraX, UCSF Chimera and VMD.</strong> PyMOL prices are per year for a one-year term, in US dollars, from the vendor's price list. Entries checked between November 10 and 12, 2026. Sources: Schrödinger (2026), UCSF RBVI (2026), University of Illinois (2026).</caption>
  <thead>
    <tr>
      <th scope="col"></th>
      <th scope="col">PyMOL</th>
      <th scope="col">ChimeraX</th>
      <th scope="col">UCSF Chimera</th>
      <th scope="col">VMD</th>
    </tr>
  </thead>
  <tbody>
    <tr>
      <td>Academic or non-profit</td>
      <td>\$99 single user, \$500 for a lab or department</td>
      <td>Free</td>
      <td>Free</td>
      <td>Free, registration to download</td>
    </tr>
    <tr>
      <td>Student or teacher</td>
      <td>No cost under the educational subscription</td>
      <td>Free</td>
      <td>Free</td>
      <td>Free</td>
    </tr>
    <tr>
      <td>Government</td>
      <td>\$315 single user, \$1,575 lab or department</td>
      <td>Free</td>
      <td>Free</td>
      <td>Free for research use</td>
    </tr>
    <tr>
      <td>Company</td>
      <td>\$1,050 single user (25 or fewer employees), \$5,250 under 1,000 employees</td>
      <td>Separate written licence; no price published on the pages we read</td>
      <td>Licence cost depends on the number of users</td>
      <td>Internal business use allowed; commercial use needs a commercial licence</td>
    </tr>
    <tr>
      <td>Source code</td>
      <td>Open-source build on GitHub, BSD-like licence</td>
      <td>Non-commercial licence</td>
      <td>Non-commercial licence</td>
      <td>Included, but not under an open-source licence</td>
    </tr>
    <tr>
      <td>Platforms</td>
      <td>64-bit Windows 11, macOS 13 or later including Apple Silicon, Linux (glibc 2.28 or later)</td>
      <td>Windows, macOS and some Linux distributions; not supported in virtual machines</td>
      <td>64-bit Windows, macOS and Linux</td>
      <td>1.9.4 for macOS, Windows and Unix; 2.0.0 for Unix only</td>
    </tr>
    <tr>
      <td>Current version</td>
      <td>3.1.9, updated October 7, 2026</td>
      <td>1.12, production release June 11, 2026</td>
      <td>1.20, June 25, 2026 (legacy)</td>
      <td>2.0.0, March 25, 2026, listed as alpha on the home page</td>
    </tr>
  </tbody>

Two cautions on the table. PyMOL's two pages disagree slightly on who qualifies for the free educational build, and that build is "not provided for the purposes of academic research or publication". ChimeraX being free does not make it open source: the licence restricts it to non-commercial purposes.

VMD's version depends on which page you read: the download page lists release 2.0.0 for 64-bit Linux, while the home page calls 2.0.0 an alpha and still offers 1.9.4 for macOS and Windows.

## How ChimeraX replaced UCSF Chimera

UCSF Chimera is the program ChimeraX replaced. The Chimera home page says it is no longer under active development and urges users to migrate, and its download page recommends ChimeraX "for higher performance and many new features". The last production release, 1.20, appeared on June 25, 2026 and only fixed problems on Mac OS Tahoe. Daily builds from October 2026 are untested builds rather than releases, and 32-bit releases are no longer supported.

ChimeraX is a new program with a different command language, not an upgrade. Chimera writes chains after residues (`:25-27.A@ca`) and numbers models from `#0`; ChimeraX puts the chain first (`/A:25-27@CA`) and numbers models from `#1`. Some commands kept their names and behaviour: `matchmaker` in both builds a pairwise sequence alignment and then fits one atom per residue, C-alpha in proteins and C4' in nucleic acids.

Sessions do not carry over directly. The route Chimera documents is its ChimeraX export, which writes a Python file that rebuilds the scene in ChimeraX, and the export page lists what it leaves out, such as per-vertex surface colourings and nucleotide representations. Chimera is still free for non-commercial use, and commercial licensing costs depend on the number of users.

The usual reason to search "chimera vs chimerax" is an old tutorial or methods section that names Chimera. Follow it in ChimeraX with the commands translated, since ChimeraX's authors describe it as the next-generation program from the same group. A search for "chimera alternative" is almost always asking for ChimeraX, and "chimerax download" should go to the [UCSF ChimeraX download page](https://www.cgl.ucsf.edu/chimerax/download.html).

## VMD for trajectories, PyMOL for figures

VMD (Visual Molecular Dynamics) was built to show simulations, and its file support shows it. Its molfile plugins read CHARMM and NAMD DCD, GROMACS XTC, TRR and TNG, Amber NetCDF and LAMMPS files, and DCD is the one trajectory format on that list it also writes. Its plugin list includes analysis tools such as MultiSeq, which performs multiple sequence and structure alignments, and the program is scripted in both Tcl and Python.

PyMOL can play a trajectory too. `load_traj` uses the same molfile plugin backend as VMD, and its wiki page shows DCD and XTC examples, though it was last edited in 2019 and lists no formats. ChimeraX opens DCD, XTC, TRR, Amber NetCDF and LAMMPS dump files over a structure already open. What separates VMD is the analysis around the playback, not the playback itself.

The way you build a picture also differs. In VMD each molecule carries representations, each with its own atom selection, colouring method and drawing style, and the text commands change those settings and add a representation from them. PyMOL attaches colours and styles to selections directly, which is why `color cyan, chain A` is one line in PyMOL and three in VMD (Table 2).

For figures, PyMOL's `ray` runs its own ray tracer. VMD's `render` command offers several renderers, including an internal Tachyon ray tracer, and `render snapshot` captures the window as it is, written as a Targa file on Unix and a Windows bitmap on Windows.

A lab that publishes structure figures and rarely simulates has little reason to leave PyMOL. A lab that runs simulations every week will end up in VMD or a Python library such as MDAnalysis, and the [VMD alternatives for playing an XTC or DCD trajectory](/guides/vmd-alternatives) include a browser option for collaborators who have nothing installed.

## The same six tasks in all four

The command languages are the clearest difference. PyMOL takes comma-separated arguments and a selection language written as words (`chain A and resi 25-27`). ChimeraX and Chimera take space-separated arguments and atom specifications built from punctuation, in different orders. VMD is a Tcl shell: selections are objects you create with `atomselect` and then pass to other commands.

Table 2 runs one workflow on one structure in each. The structure is [1HSG](https://www.rcsb.org/structure/1HSG), HIV-1 protease bound to the inhibitor indinavir, with [3OXC](https://www.rcsb.org/structure/3OXC), the same enzyme with saquinavir, as the second structure to align.

  <caption><strong>Table 2. Six everyday tasks as PyMOL, ChimeraX, Chimera and VMD commands.</strong> Commands follow each program's documentation. 1HSG has two copies of the protease, chains A and B, and residues 25 to 27 hold the catalytic Asp-Thr-Gly. Models are numbered in the order opened: 1HSG is #1 in ChimeraX, #0 in Chimera and molecule 0 in VMD. ColorID 10 is cyan in VMD's colour table. In the VMD alignment, `$ca1` and `$ca2` stand for C-alpha selections of the same residues in molecules 0 and 1. Sources: PyMOL Wiki (2026), UCSF ChimeraX User Guide (2026), UCSF Chimera User's Guide (2026), VMD User's Guide (2016).</caption>
  <thead>
    <tr>
      <th scope="col">Task</th>
      <th scope="col">PyMOL</th>
      <th scope="col">ChimeraX</th>
      <th scope="col">Chimera</th>
      <th scope="col">VMD</th>
    </tr>
  </thead>
  <tbody>
    <tr>
      <td>Open a structure by PDB ID</td>
      <td>`fetch 1hsg`</td>
      <td>`open 1hsg`</td>
      <td>`open 1hsg`</td>
      <td>`mol pdbload 1hsg`</td>
    </tr>
    <tr>
      <td>Colour one chain</td>
      <td>`color cyan, chain A`</td>
      <td>`color /A cyan`</td>
      <td>`color cyan :.A`</td>
      <td>`mol selection "chain A"`, `mol color ColorID 10`, `mol addrep top`</td>
    </tr>
    <tr>
      <td>Select residues</td>
      <td>`select site, chain A and resi 25-27`</td>
      <td>`select /A:25-27`</td>
      <td>`select :25-27.A`</td>
      <td>`set site [atomselect top "chain A and resid 25 to 27"]`</td>
    </tr>
    <tr>
      <td>Measure a distance</td>
      <td>`distance d1, chain A and resi 25 and name CA, chain B and resi 25 and name CA`</td>
      <td>`distance /A:25@CA /B:25@CA`</td>
      <td>`distance :25.A@ca :25.B@ca`</td>
      <td>Select each atom as `$a` and `$b`, then `vecdist [measure center $a] [measure center $b]`</td>
    </tr>
    <tr>
      <td>Align a second structure</td>
      <td>`align 3oxc, 1hsg`</td>
      <td>`matchmaker #2 to #1`</td>
      <td>`matchmaker #0 #1`</td>
      <td>`set M [measure fit $ca2 $ca1]`, then `[atomselect 1 all] move $M`</td>
    </tr>
    <tr>
      <td>Save an image</td>
      <td>`png 1hsg.png, width=2000, dpi=300, ray=1`</td>
      <td>`save 1hsg.png width 2000 supersample 3`</td>
      <td>`copy file 1hsg.png width 2000 supersample 3`</td>
      <td>`render snapshot 1hsg.tga`</td>
    </tr>
  </tbody>

Each row hides a difference in behaviour. PyMOL's `fetch` takes a `type` argument for maps and other formats, ChimeraX's `open` treats a bare four-character identifier as an RCSB PDB entry, Chimera fetches any input that resembles a PDB ID, and VMD's `mol pdbload` retrieves the file from the RCSB site.

Alignment is where the four really part. PyMOL's `align` pairs residues by sequence and rejects outliers over several cycles, then prints the RMSD. ChimeraX's `matchmaker` fits C-alpha atoms iteratively with a 2.0 Å cutoff and reports the RMSD over both the pruned pairs and all pairs, and Chimera's takes the reference first and the structure to move second.

VMD does none of the pairing for you. `measure fit` returns the matrix that minimizes the weighted RMSD between two selections, and the user guide's own example moves the whole molecule with it, not only the fitted atoms.

So `protein and name CA` works only when both entries contain exactly the same residues; otherwise you name the residue range in each selection. The programs can therefore give different [RMSD](/guides/rmsd) values for the same pair, and none of them is wrong, because they describe different subsets of atoms.

The image commands differ in what they promise. PyMOL's `ray=1` runs its ray tracer, which can take several minutes on a complex scene, and `dpi` sets the file's resolution metadata. ChimeraX and Chimera render what the window shows, supersampled, at a width you give in pixels, and Chimera's `copy` accepts supersampling from 1 to 4. VMD's snapshot is a screen grab; use one of its `render` ray tracers for a publication figure.

## AlphaFold models and density maps

ChimeraX has AlphaFold built in. `alphafold fetch p29474` downloads the database model for a UniProt name or accession, `alphafold match #1` finds the models that match the chains of an open experimental structure and reports the RMSD for each, and `alphafold pae` draws the predicted aligned error plot. The confidence values it shows are [pLDDT](/guides/plddt) per residue and [PAE](/guides/pae) between residues, and the ChimeraX tutorials include a walkthrough on AlphaFold predictions.

PyMOL opens AlphaFold models as ordinary structure files. Its wiki lists no AlphaFold identifier among the types `fetch` accepts, so the usual route is to download the file from the AlphaFold Database, `load` it, and colour by B-factor, because the pLDDT confidence sits in that column. Chimera and VMD take the same route; we did not find AlphaFold commands in either program's documentation.

Density maps follow the same split. ChimeraX opens an EMDB map directly (`open emdb:3087` appears in its documentation) and treats maps as first-class objects alongside atomic models. PyMOL's `fetch` can download X-ray maps for a PDB entry with `type=2fofc` or `type=fofc`, as in `fetch 1cll, type=2fofc`.

What each program does after that, such as fitting models into cryo-EM maps or rebuilding, is where the ChimeraX Toolshed matters, and we did not benchmark it. All of them can display the [electron density maps](/guides/electron-density-map) that [X-ray crystallography and cryo-EM](/guides/cryo-em-vs-x-ray-nmr-alphafold) produce, and none can tell you by itself whether a ligand really fits the density.

## Plugins and scripting

All four are scriptable. PyMOL runs `.pml` command scripts and has a Python API (`cmd.align(...)`), and its Plugin Manager installs plugins from a Python file, a zip or a wiki page URL. The same wiki page notes that plugins written in Tkinter are deprecated, with removal expected by PyMOL 4.0, so older plugins may need porting to Qt.

ChimeraX's `runscript` runs a Python script (`.py`) or a ChimeraX command file (`.cxc`), and a `.cxc` file can be applied to a list of files with `open myscript.cxc foreach ~/data/*.cif`. Third-party extensions come as bundles from the Toolshed, whose tag counts list 23 in structure analysis and 23 in EM, with ISOLDE and Clipper among the examples. Tags overlap, so those are not a bundle total.

VMD's text console is Tcl, with a Python interface alongside, and its plugins cover analysis, modelling and file formats. The Tcl selections in Table 2 are the same objects a script loops over frame by frame, which is why VMD scripts are common in MD methods sections. Chimera's scripts and extensions belong to the legacy program and need rewriting for ChimeraX.

In practice the difference is history. A lab with years of PyMOL scripts and a [PyMOL command cheat sheet](/guides/pymol-commands) pinned above the bench has a strong reason to stay, a group doing cryo-EM model building has a strong reason to use ChimeraX, and a simulation group already has VMD scripts in its repository.

## Large complexes

The one performance statement we can source is the Chimera download page's, which recommends ChimeraX "for higher performance".

We found no published head-to-head timing of PyMOL, ChimeraX and VMD on the same large assembly, so we do not claim one is faster for a ribosome or a viral capsid. All are desktop programs that use the graphics card, and ChimeraX's documentation says it is not supported in virtual machines or over remote display. If a structure feels slow, test it in each on your own machine, since GPU drivers matter more than the program for this question.

## Running PyMOL and ChimeraX commands in the PDB viewer

The [PDB viewer](/app/pdb-viewer) by ProteinIQ runs in the browser and accepts PyMOL commands, scripts and session files, along with a subset of ChimeraX commands. That subset includes `matchmaker` (also `mm`, including `matchmaker all to 1hsg/A`), `color byattribute`, `split`, `combine`, `torsion`, `clip`, `clashes` and `contacts`. It also plays XTC, DCD and TRR trajectories over a PDB or GRO topology.

A command outside what the viewer implements answers "not supported yet": PyMOL's `ray` is one, VMD's Tcl commands and Chimera's atom specifications are not read, and ChimeraX bundles and the Toolshed have no counterpart. It is a way to try the PyMOL and ChimeraX command styles on a structure, not a replacement for any of the four programs.

## Which one to pick

  <caption><strong>Table 3. Which program fits which situation.</strong> A summary of the comparison above, from the sources cited there.</caption>
  <thead>
    <tr>
      <th scope="col">Your situation</th>
      <th scope="col">Start with</th>
      <th scope="col">Check</th>
    </tr>
  </thead>
  <tbody>
    <tr>
      <td>Academic lab, no budget, new to structure viewing</td>
      <td>ChimeraX</td>
      <td>Whether an industry sponsor makes the work commercial under the licence</td>
    </tr>
    <tr>
      <td>Lab with existing PyMOL scripts and figures</td>
      <td>PyMOL</td>
      <td>That the lab holds a current licence, or uses the open-source build</td>
    </tr>
    <tr>
      <td>MD trajectories (DCD, XTC, TRR) and their analysis</td>
      <td>VMD</td>
      <td>That your build exists for your platform, and the licence if you work for a company</td>
    </tr>
    <tr>
      <td>Company or industry-funded research</td>
      <td>PyMOL (priced), ChimeraX or VMD (ask the university)</td>
      <td>The written commercial licence before any use</td>
    </tr>
    <tr>
      <td>Cryo-EM maps or AlphaFold models</td>
      <td>ChimeraX</td>
      <td>That the Toolshed bundle you need exists</td>
    </tr>
    <tr>
      <td>A class or a one-off figure</td>
      <td>Either; free educational PyMOL or ChimeraX</td>
      <td>Educational PyMOL is not for research or publication</td>
    </tr>
    <tr>
      <td>An old tutorial that says Chimera</td>
      <td>ChimeraX, with a translated tutorial</td>
      <td>That the commands differ: ChimeraX is a rewrite</td>
    </tr>
  </tbody>

If you cannot decide between PyMOL and ChimeraX, open the same structure in both for an hour; the commands in Table 2 are enough to tell which language you prefer. If you simulate, learn VMD alongside whichever one you pick, because the figure program and the trajectory program are often different. The [list of protein visualization software](/guides/protein-visualization-software) and the [free PyMOL alternatives](/guides/pymol-alternatives) cover the other options.

## Viewing and comparing tools on ProteinIQ

Open the structure or trajectory in the viewer first, then use the other tools when you need a number or a file format.

  <caption><strong>Table 4. Structure viewing and comparison tools on ProteinIQ.</strong></caption>
  <thead>
    <tr>
      <th scope="col">Task</th>
      <th scope="col">Tool</th>
    </tr>
  </thead>
  <tbody>
    <tr>
      <td>View a structure and run PyMOL commands in the browser</td>
      <td>[PDB viewer](/app/pdb-viewer)</td>
    </tr>
    <tr>
      <td>Download a PDB entry</td>
      <td>[PDB download](/app/pdb-download)</td>
    </tr>
    <tr>
      <td>Fetch an AlphaFold database model</td>
      <td>[AlphaFold DB download](/app/alphafold-database-download)</td>
    </tr>
    <tr>
      <td>Calculate the RMSD between two structures</td>
      <td>[RMSD calculator](/app/rmsd-calculator)</td>
    </tr>
    <tr>
      <td>Analyse an MD trajectory on a server</td>
      <td>[MD trajectory analysis](/app/md-trajectory-analysis)</td>
    </tr>
    <tr>
      <td>Compute an electrostatic surface</td>
      <td>[PDB2PQR](/app/pdb2pqr)</td>
    </tr>
  </tbody>

## Frequently asked questions

### Is ChimeraX free for companies?

No. ChimeraX is free for academic, government, nonprofit and personal use, and commercial use needs a separately executed written licence from UCSF. The licence also excludes research funded by an entity that commercially exploits similar software, so industry-funded academic work needs a check.

### What is the difference between Chimera and ChimeraX?

ChimeraX is the next-generation rewrite of UCSF Chimera from the same group, with a different command language: Chimera writes `:25-27.A` where ChimeraX writes `/A:25-27`. Chimera is legacy software; its last release, 1.20, appeared on June 25, 2026, and the site recommends moving to ChimeraX.

### Is PyMOL free?

Partly. Open-source PyMOL is published on GitHub under a BSD-like licence, and full-time students and their educators can get a free educational build, but the installers on pymol.org are a licensed product. An academic single-user licence is \$99 per year.

### Is VMD better than PyMOL?

For molecular dynamics, usually. VMD reads DCD, XTC, TRR, TNG and Amber NetCDF trajectories and comes with analysis plugins, while PyMOL plays trajectories through the same file readers but is mainly a figure program. VMD is free for academic, research and internal business use; commercial use needs a commercial licence.

### Can ChimeraX open AlphaFold models?

Yes. `alphafold fetch` downloads a model by UniProt name or accession, `alphafold match` finds models for the chains of an open structure, and `alphafold pae` plots the predicted aligned error.

### What does matchmaker do in ChimeraX?

It superimposes one structure on another by sequence alignment and then fits C-alpha atoms iteratively with a 2.0 Å cutoff, and reports the final RMSD over all residue pairs and over the pruned set. Chimera's `matchmaker` works the same way, and PyMOL's nearest equivalents are `align` and `super`.

### Where can I learn ChimeraX?

The ChimeraX site lists tutorials with click-to-execute commands, including a quick start guide, protein-ligand binding sites, and AlphaFold analysis. Open them inside ChimeraX so the links run.
