Json-Performance

September 3, 2026 · View on GitHub

Performance profiling of JSON libraries (Compiled and run on Linux 6.18.33.2-microsoft-standard-WSL2 using the GCC 16.1.0 compiler).

Latest Results: (Sep 02, 2026)

Using the following commits:


| Jsonifier: a1bd83c
| Glaze: a5f515c
| Simdjson: 06856ec

Active Implementations:

LibraryActive Implementation
JsonifierAVX2
simdjson (ondemand)haswell
Glaze (utf8-validation)AVX2
Glaze (string-escape)AVX2
Glaze (float-write)SSE4.1

Each library selects its own instruction-set implementation at build or run time; the values above are the implementations that produced the results below. Glaze reports per-subsystem backends, which may differ from one another within a single build.

Adaptive sampling on (Intel(R) Core(TM) i9-14900KF-AVX2): iterations begin at 80 and double each epoch (e.g. 80 → 160 → 320 → ...) up to a maximum of 100000 iterations. Each epoch runs all iterations and evaluates a trailing window of max(iterations/10, 30) samples, capped at 100000. Convergence requires RSE < 2.5% AND mean shift < 1% epoch-over-epoch simultaneously. The first epoch satisfying both conditions is retained as the canonical result. If convergence is never reached before 10 seconds elapse or the iteration cap is hit, the result is marked non-converged and excluded from all rankings — only converged results participate in win/tie/loss tallying. All results use Bessel-corrected variance and Welch's t-test for statistical tie detection.

(All of the libraries are performing UTF8-validation in these tests. Jsonifier is only performing "structural indexing/stage-1 + stage-2" parsing for the 'partial' tests here, for the rest of them - we perform scalar structural iteration)

Note:

This is the commit of BenchmarkSuite that was used to generate these results: f0ef39c.


Double Test Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier1219.30.1765271.21526ms181130187.5681416.472.453781(Win)
glaze875.3250.2342060.745446ms181130640.6451973.13.432362(Loss)
simdjson (ondemand)229.3840.1101112.08656ms1811302062.017529.313.20673(Loss)

Double Test Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier721.7820.08387260.862974ms181130120.8332392.834.153621(Win)
glaze493.190.2287881.10814ms1798301898.193476.776.097462(Loss)

Uint64 Test Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier1998.010.04764650.742762ms38733023.27471848.631.497791(Win)
glaze1863.060.04059330.754219ms38733019.42991982.531.609482(Loss)
simdjson (ondemand)456.0650.2457512.26283ms38733011883.88098.86.641793(Loss)

Uint64 Test Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier3033.240.08041950.608492ms38733028.7691217.70.9749981(Win)
glaze1408.030.1769410.967717ms387330646.3232623.232.136332(Loss)

Int64 Test Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier2067.330.03802980.723957ms38623013.77131781.571.449471(Win)
glaze1721.230.05683590.777276ms38623044.37242139.81.749392(Loss)
simdjson (ondemand)463.2581.619752.17105ms3862304975017950.46.537213(Loss)

Int64 Test Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier2930.450.2587940.588545ms386230317.3851256.831.010341(Win)
glaze1405.670.1040320.95099ms386230222.9022620.172.142492(Loss)

Bool Test Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier1833.990.03470020.385191ms905300.8470.61.59051(Win)
glaze643.3960.03987841.10328ms905329.157261341.444.658812(Loss)
simdjson (ondemand)198.9250.09535261.43361ms90530513.4594338.715.20223(Loss)

Bool Test Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier1305.970.3044310.414043ms90530121.43660.8672.262691(Win)
glaze344.9170.1259240.922651ms90530297.8572502.278.730612(Loss)

String Test Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier3705.090.03497080.929425ms95783022.29892465.330.8111331(Win)
glaze2964.590.1540621.04172ms957830675.9823081.131.015132(Loss)
simdjson (ondemand)1491.980.0627591.81858ms957830442.8926122.272.029983(Loss)

String Test Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier5981.20.1712740.635928ms957830205.2471527.170.4982981(Win)
glaze2954.140.10941.03573ms957830343.2753092.031.016842(Loss)

Abc (In Order) Partial Test (Minified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier5290.160.89778314.5437ms233995304.30268e+06421830.5741531(Win)
simdjson (ondemand)3586.051.1526314.56ms233995301.54342e+0762228.60.8471592(Loss)
glaze2150.70.529627.0014ms233995309.0588e+061037591.412783(Loss)

Abc (In Order) Partial Test (Prettified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier7014.280.51376811.9993ms346753301.76007e+0647145.20.4330591(Win)
simdjson (ondemand)5249.151.4170115.9466ms346753302.39072e+0762998.70.5787552(Loss)
glaze2731.520.48897230.1473ms346753301.05129e+071210641.112433(Loss)

Abc (In Order) Test (Minified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier1856.680.64891129.7757ms233995301.82487e+071201901.636531(Win)
glaze1718.061.4434332.4407ms233995301.05451e+081298871.768382(Loss)
simdjson (ondemand)1559.110.75836836.1196ms233995303.53463e+071431301.948783(Loss)

Abc (In Order) Test (Minified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier4495.850.34816512.5184ms2339953089594349635.70.6749991(Win)
glaze4179.750.91130213.493ms233995307.10164e+0653389.60.7266792(Loss)

Abc (In Order) Test (Prettified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
glaze STATISTICAL TIE2142.641.0337840.2389ms346753307.63694e+071543381.418171(Tie)
simdjson (ondemand) STATISTICAL TIE2134.251.1796138.4889ms346753301.00218e+081549441.423671(Tie)
jsonifier1592.020.587081122.113ms346753324.75874e+072077171.908753(Loss)

Abc (In Order) Test (Prettified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier4785.121.43596172.575ms3467531281.26052e+0869107.80.6344141(Win)
glaze4456.80.55830219.6976ms346753305.14818e+0674198.90.6816022(Loss)

Abc (Out of Order) Partial Test (Minified) Read Results (View the data used in the following test):

The JSON documents in these two tests feature keys ranging from "a" to "z", where each key corresponds to an array of values. Notably, the library parsing methods in this test arrange these keys in reverse order, deviating from the typical "a" to "z" arrangement. This test effectively demonstrates the challenges encountered when utilizing simdjson and iterative parsers that lack the ability to efficiently locate memory locations through hashing. In cases where the keys are not in the expected sequence, performance is significantly compromised, with the severity escalating as the document size increases. In contrast, hash-based solutions offer a viable alternative by circumventing these issues and maintaining optimal performance regardless of the JSON document's scale, or ordering of the keys being parsed.

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier5218.520.16413110.7723ms2339953014778142762.10.5820831(Win)
simdjson (ondemand)2771.460.37048819.501ms233995302.66973e+0680519.11.096252(Loss)
glaze2161.210.43962124.9597ms233995306.18153e+061032541.405913(Loss)

Abc (Out of Order) Partial Test (Prettified) Read Results (View the data used in the following test):

The JSON documents in these two tests feature keys ranging from "a" to "z", where each key corresponds to an array of values. Notably, the library parsing methods in this test arrange these keys in reverse order, deviating from the typical "a" to "z" arrangement. This test effectively demonstrates the challenges encountered when utilizing simdjson and iterative parsers that lack the ability to efficiently locate memory locations through hashing. In cases where the keys are not in the expected sequence, performance is significantly compromised, with the severity escalating as the document size increases. In contrast, hash-based solutions offer a viable alternative by circumventing these issues and maintaining optimal performance regardless of the JSON document's scale, or ordering of the keys being parsed.

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier6638.20.58972812.2509ms346753302.58919e+0649816.10.4575731(Win)
simdjson (ondemand)3896.790.85093520.7727ms346753301.56437e+07848620.7796792(Loss)
glaze2791.830.34270429.0321ms346753304.94337e+061184491.08843(Loss)

Abc (Out of Order) Test (Minified) Read Results (View the data used in the following test):

The JSON documents in these two tests feature keys ranging from "a" to "z", where each key corresponds to an array of values. Notably, the library parsing methods in this test arrange these keys in reverse order, deviating from the typical "a" to "z" arrangement. This test effectively demonstrates the challenges encountered when utilizing simdjson and iterative parsers that lack the ability to efficiently locate memory locations through hashing. In cases where the keys are not in the expected sequence, performance is significantly compromised, with the severity escalating as the document size increases. In contrast, hash-based solutions offer a viable alternative by circumventing these issues and maintaining optimal performance regardless of the JSON document's scale, or ordering of the keys being parsed.

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
glaze1859.050.7100428.9844ms233995302.1793e+071200371.634491(Win)
jsonifier1654.360.35965133.2093ms233995307.06051e+061348891.836782(Loss)
simdjson (ondemand)280.0580.897339187.929ms233995301.53374e+0979681610.8523(Loss)

Abc (Out of Order) Test (Minified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier4805.710.68387711.6566ms233995303.02535e+0646435.40.6315931(Win)
glaze4402.320.47425412.7781ms233995301.73377e+0650690.30.6899572(Loss)

Abc (Out of Order) Test (Prettified) Read Results (View the data used in the following test):

The JSON documents in these two tests feature keys ranging from "a" to "z", where each key corresponds to an array of values. Notably, the library parsing methods in this test arrange these keys in reverse order, deviating from the typical "a" to "z" arrangement. This test effectively demonstrates the challenges encountered when utilizing simdjson and iterative parsers that lack the ability to efficiently locate memory locations through hashing. In cases where the keys are not in the expected sequence, performance is significantly compromised, with the severity escalating as the document size increases. In contrast, hash-based solutions offer a viable alternative by circumventing these issues and maintaining optimal performance regardless of the JSON document's scale, or ordering of the keys being parsed.

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
glaze2105.061.5886539.9581ms346753301.86849e+081570931.443481(Win)
jsonifier1550.952.4453651.4555ms346753308.15561e+082132181.959142(Loss)
simdjson (ondemand)417.8680.497477190.531ms346753304.64975e+087913727.273063(Loss)

Abc (Out of Order) Test (Prettified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier5071.180.7775917.5894ms346753307.71338e+0665209.50.5988751(Win)
glaze4044.261.217620.0348ms346753302.97365e+0781767.60.7505932(Loss)

Apache Builds Test (Minified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier1306.880.51614817.3132ms94651303.81283e+0669069.92.324031(Win)
glaze1286.320.5889717.0578ms94651305.12457e+0670173.82.361782(Loss)
simdjson (ondemand)1248.631.3174718.2672ms94651302.72139e+0772292.52.432983(Loss)

Apache Builds Test (Minified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
glaze STATISTICAL TIE5734.981.298264.12885ms94651301.25265e+0615739.60.5288731(Tie)
jsonifier STATISTICAL TIE5558.581.2924.2793ms94651301.32059e+0616239.10.5453831(Tie)

Apache Builds Test (Prettified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
simdjson (ondemand)1668.161.242745.601ms136024322.98838e+0777763.71.821081(Win)
glaze1489.90.70126421.1514ms136024301.11842e+0787068.22.038612(Loss)
jsonifier1193.470.73043926.636ms136024301.89103e+071086942.545883(Loss)

Apache Builds Test (Prettified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier6508.42.1764699.352ms1360242564.81753e+0719931.60.466141(Win)
glaze5474.740.9678516.17386ms136046301.57827e+0623698.60.5544192(Loss)

Canada Test (Minified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier787.060.244479607.239ms2090234301.15021e+092.53272e+063.861191(Win)
simdjson (ondemand)760.2210.246929643.745ms2090234301.25769e+092.62213e+063.99732(Loss)
glaze667.2621.95517689.312ms2090234301.0235e+112.98744e+064.554213(Loss)

Canada Test (Minified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier1524.71.20247314.689ms2090234307.4147e+091.30741e+061.992821(Win)
glaze1329.871.31871355.385ms2090234301.17216e+101.49894e+062.284792(Loss)

Canada Test (Prettified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
simdjson (ondemand)2121.950.970392705.163ms6661897302.53245e+102.99408e+061.432131(Win)
jsonifier1850.871.31914822.067ms6661897306.15098e+103.43258e+061.64192(Loss)
glaze1662.141.22225919.073ms6661897306.54785e+103.82235e+061.828373(Loss)

Canada Test (Prettified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
glaze3102.290.482593511.812ms6661897302.93032e+092.04793e+060.9794281(Win)
jsonifier3037.870.435681506.973ms6661897302.49066e+092.09136e+061.000242(Loss)

CitmCatalog Test (Minified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier1876.790.27435661.8782ms500299301.45942e+072542231.619261(Win)
glaze1788.860.42978164.5157ms500299303.94207e+072667191.698862(Loss)
simdjson (ondemand)1152.830.438846521.012ms500299642.11123e+084138722.6363(Loss)

CitmCatalog Test (Minified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier STATISTICAL TIE5460.210.49494522.3958ms500299305.61145e+0687381.60.5562661(Tie)
glaze STATISTICAL TIE5265.891.758452.8896ms500299328.12269e+0790606.20.5766041(Tie)

CitmCatalog Test (Prettified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
simdjson (ondemand)2919.940.491728112.032ms1439562301.60356e+084701721.040681(Win)
jsonifier2855.760.688477116.905ms1439562303.28638e+084807381.064112(Loss)
glaze2794.270.694646120.985ms1439562303.4944e+084913171.087523(Loss)

CitmCatalog Test (Prettified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier7075.531.4250248.4611ms1439562302.29354e+081940310.4291331(Win)
glaze4623.932.25455167.215ms1439584321.43391e+092969110.6569072(Loss)

Discord Test (Minified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier1910.410.04072137.50247ms56369303939.0728139.41.589341(Win)
glaze1742.990.5547587.57398ms563693087825730842.31.742032(Loss)
simdjson (ondemand)1269.411.181210.9826ms56369307.50662e+0642348.52.39263(Loss)

Discord Test (Minified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
glaze6364.160.5173552.34076ms563693057292.38446.930.4759311(Win)
jsonifier5413.880.1411542.81796ms56369305893.499929.60.5575592(Loss)

Discord Test (Prettified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
glaze2308.980.890059.73028ms94370303.61058e+0638977.51.315281(Win)
simdjson (ondemand)2070.160.77247132.7539ms94370323.6089e+0643474.11.467122(Loss)
jsonifier1851.41.3895712.3505ms94370301.36883e+07486111.640663(Loss)

Discord Test (Prettified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier7095.960.04825173.56975ms94370301123.55126830.4275081(Win)
glaze5387.781.5716444.8035ms943701288.82198e+0616704.20.5630122(Loss)

Github Events Test (Minified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
simdjson (ondemand)1637.561.510747.9279ms52708306.45148e+0630695.91.854151(Win)
jsonifier1466.81.145568.91875ms52708304.62348e+0634269.42.070132(Loss)
glaze1289.681.598499.95347ms52708301.16448e+0738975.82.354583(Loss)

Github Events Test (Minified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
glaze7201.090.6948042.35282ms527083070567.16980.370.4201731(Win)
jsonifier6049.182.369985.11747ms52708321.24107e+068309.590.4980642(Loss)

Github Events Test (Prettified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
simdjson (ondemand)2220.31.258367.98568ms70103304.30703e+06301111.367421(Win)
glaze1568.041.583810.7455ms70103301.36798e+0742636.31.936722(Loss)
jsonifier1319.570.95413214.3694ms70103307.01051e+0650664.72.301983(Loss)

Github Events Test (Prettified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier6393.730.1432053.01104ms70103306726.7310456.40.47191(Win)
glaze5693.50.86470928.8169ms701031281.31967e+0611742.40.5322542(Loss)

Google Maps Response Test (Minified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier1529.280.9019732.13676ms11812301324297366.11.981181(Win)
glaze1378.280.1369035.08258ms11812324006.358173.092.197892(Loss)
simdjson (ondemand)1085.940.7910092.87127ms118123020198610373.32.791783(Loss)

Google Maps Response Test (Minified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier5936.560.5181011.47659ms11812323092.841897.530.5015111(Win)
glaze4738.220.1676151.05039ms1181230476.3922377.430.633132(Loss)

Google Maps Response Test (Prettified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier STATISTICAL TIE2682.61.930086.66893ms31235321.46984e+0611104.21.130291(Tie)
simdjson (ondemand) STATISTICAL TIE2631.670.06302223.12765ms31235301526.6211319.11.152291(Tie)
glaze2214.282.100423.57944ms31235302.39524e+0613452.71.36973(Loss)

Google Maps Response Test (Prettified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier8222.60.1205851.26283ms3123530572.4933622.70.3628641(Win)
glaze6706.340.2143851.51199ms31235302720.324441.770.4500542(Loss)

Instruments Test (Minified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier3717.620.18626616.7821ms108313328571327785.30.8166851(Win)
glaze2142.411.2044312.3319ms108313301.01167e+0748214.51.417782(Loss)
simdjson (ondemand)1612.671.8166415.8654ms108313304.06189e+0764052.41.883773(Loss)

Instruments Test (Minified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier6582.90.08497214.22677ms108313305333.3615691.50.4603711(Win)
glaze5816.661.034814.67788ms108313301.0131e+0617758.50.5215312(Loss)

Instruments Test (Prettified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
simdjson (ondemand)2960.210.75292616.9617ms213963308.08087e+0668931.21.026251(Win)
glaze2843.510.9616817.7936ms213963301.42873e+0771760.21.068432(Loss)
jsonifier2328.650.88974823.4401ms213963301.82358e+0787626.41.304713(Loss)

Instruments Test (Prettified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier8453.890.07505556.30405ms213963309845.7924136.90.3587591(Win)
glaze5857.730.9282359.04076ms213963303.13658e+0634834.50.5184112(Loss)

Marine IK Test (Minified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier805.50.299421528.504ms1834197301.26837e+092.1716e+063.772421(Win)
simdjson (ondemand) STATISTICAL TIE587.7780.288747716.589ms1834197302.21524e+092.976e+065.169892(Tie)
glaze STATISTICAL TIE578.2531.20708708.051ms1834197303.99987e+103.02502e+065.255232(Tie)

Marine IK Test (Minified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier1007.130.256455416.649ms1834197305.95207e+081.73685e+063.017141(Win)
glaze923.3570.672478458.73ms1833577304.8656e+091.89378e+063.290872(Loss)

Marine IK Test (Prettified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
simdjson (ondemand)2744.40.382356821.44ms9930848305.22317e+093.45095e+061.107281(Win)
glaze2441.790.230014929.309ms9930848302.38774e+093.87863e+061.244542(Loss)
jsonifier2346.990.359808974.463ms9930848306.32429e+094.03529e+061.294823(Loss)

Marine IK Test (Prettified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier2583.272.33022821.118ms9930848302.18953e+113.66621e+061.17611(Win)
glaze2024.651.426981050.32ms9930228301.33652e+114.67745e+061.500642(Loss)

Mesh Test (Minified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier1409.720.288679105.651ms642697304.72605e+074347832.15571(Win)
simdjson (ondemand)1174.220.545424125.239ms642697302.43168e+085219852.587952(Loss)
glaze1049.660.560211139.925ms642697303.21028e+085839282.895193(Loss)

Mesh Test (Minified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier1642.890.32320990.3613ms642697304.36197e+073730761.849631(Win)
glaze1416.840.791253103.813ms642692303.51493e+084325962.144052(Loss)

Mesh Test (Prettified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
simdjson (ondemand)2208.30.591669129.496ms1225964302.94388e+085294441.375991(Win)
glaze1784.060.3277157.834ms1225964301.38359e+086553411.703262(Loss)
jsonifier1641.370.628794169.601ms1225964306.01838e+087123131.85113(Loss)

Mesh Test (Prettified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier2553.140.609955110.335ms1225964302.34058e+084579351.189951(Win)
glaze2129.161.26758129.443ms1225970301.45349e+095491261.426772(Loss)

Random Test (Minified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier1142.10.38702183.4072ms409725305.25982e+073421292.660741(Win)
simdjson (ondemand)964.1160.71790498.3656ms409725302.53969e+084052883.151772(Loss)
glaze914.3050.4768104.137ms409725301.24565e+084273683.323773(Loss)

Random Test (Minified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier4343.262.4654321.6991ms409725301.4759e+0889965.60.6991491(Win)
glaze3589.441.1903228.0625ms409725305.03712e+071088590.8462692(Loss)

Random Test (Prettified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
simdjson (ondemand)1750.290.410249103.976ms785750309.25472e+074281281.736051(Win)
glaze1383.620.90057129.078ms785750307.13661e+085415872.196132(Loss)
jsonifier1170.50.963508152.625ms785750301.14146e+096401972.595993(Loss)

Random Test (Prettified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier5481.370.62234934.4413ms785750302.1716e+071367080.5540911(Win)
glaze3740.772.0983846.8125ms785750305.30076e+082003200.8121162(Loss)

Semanticscholar Corpus Test (Minified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier1306.590.2831851543.99ms8587914309.45277e+096.26828e+062.325921(Win)
simdjson (ondemand)1224.280.2700761665.77ms8587914309.79281e+096.6897e+062.482192(Loss)
glaze1135.141.323251684.7ms8587914302.7345e+117.21501e+062.67723(Loss)

Semanticscholar Corpus Test (Minified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier3412.621.08148578.295ms8588126302.02105e+102.4e+060.890221(Win)
glaze2979.540.502099666.471ms8588126305.71476e+092.74884e+061.019622(Loss)

Semanticscholar Corpus Test (Prettified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
simdjson (ondemand)1248.241.456721767.26ms9804437303.57212e+117.49074e+062.434541(Win)
glaze1193.810.4752761882.68ms9804437304.15709e+107.83229e+062.545662(Loss)
jsonifier1168.720.3686441973.7ms9804437302.60951e+108.00039e+062.600283(Loss)

Semanticscholar Corpus Test (Prettified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier3190.611.56936826.55ms11078090308.10113e+103.31124e+060.9521021(Win)
glaze2506.931.0392999.601ms11078090305.75392e+104.21427e+061.211852(Loss)

Twitter Partial Test (Minified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier STATISTICAL TIE5387.530.48725111.8305ms264040301.55592e+06467390.5638221(Tie)
simdjson (ondemand) STATISTICAL TIE5329.490.57228211.8157ms264040302.19336e+0647248.10.5699761(Tie)
glaze2023.150.48099630.5893ms264040301.07519e+071244631.501913(Loss)

Twitter Partial Test (Prettified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
simdjson (ondemand) STATISTICAL TIE6899.90.56243914.0628ms399947302.89996e+06552790.4402771(Tie)
jsonifier STATISTICAL TIE6852.810.46272614.8408ms399947301.98993e+0655658.80.4432881(Tie)
glaze2618.680.54554435.9774ms399947301.89419e+071456531.160383(Loss)

Twitter Test (Minified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier1714.740.360818390.625ms2640401283.59358e+071468491.772011(Win)
simdjson (ondemand)1631.390.65940738.1699ms264040303.10782e+071543521.862642(Loss)
glaze1506.831.0080641.2583ms264040308.51348e+071671112.016433(Loss)

Twitter Test (Minified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
glaze6384.821.1151110.5403ms263923305.79713e+0639421.10.4755781(Win)
jsonifier5792.22.0000811.2748ms264040302.26813e+0743473.60.5240912(Loss)

Twitter Test (Prettified) Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
simdjson (ondemand)2265.081.4629440.9149ms399947301.82061e+081683911.341481(Win)
glaze1815.732.381949.0899ms399947307.51055e+082100641.67332(Loss)
jsonifier1428.660.92913865.013ms399947301.846e+082669782.126873(Loss)

Twitter Test (Prettified) Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier6613.130.63592214.4908ms399947304.03572e+0657676.10.4593361(Win)
glaze5324.880.656338389.803ms3998302565.65492e+0771608.70.5704252(Loss)

Minify Test Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
glaze2369.090.51051345.6869ms466906302.76205e+071879531.282691(Win)
jsonifier2264.710.90849347.5365ms466906309.57194e+071966161.341692(Loss)
simdjson (ondemand)1230.211.9407383.5178ms466906301.48032e+093619532.468823(Loss)

Prettify Test Write Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
glaze3593.211.1032746.5156ms699405301.25827e+081856290.8456991(Win)
jsonifier3199.741.96217113.758ms699405325.35366e+082084560.9493332(Loss)

Validate Test Read Results (View the data used in the following test):

LibraryThroughput (MB/s)RSE (%)Window DurationFile Size (Bytes)Window Samples (k)VarianceLatency / Run (ns)Cycles/BytePosition
jsonifier3898.931.4362437.4106ms631514301.47656e+081544680.7792121(Win)
glaze2568.530.39719956.5072ms631514302.60216e+072344761.183152(Loss)