kestrelfield commited on
Commit
68945d4
·
verified ·
1 Parent(s): a10fede

Version 1.0 data, as of 13 October 2026

Browse files
Files changed (4) hide show
  1. README.md +113 -0
  2. changes.csv +0 -0
  3. estimates.csv +81 -0
  4. models.csv +0 -0
README.md ADDED
@@ -0,0 +1,113 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ pretty_name: llmledger
4
+ language:
5
+ - en
6
+ tags:
7
+ - llm
8
+ - pricing
9
+ - litellm
10
+ - survival-analysis
11
+ - competing-risks
12
+ size_categories:
13
+ - 10K<n<100K
14
+ configs:
15
+ - config_name: changes
16
+ data_files: changes.csv
17
+ default: true
18
+ - config_name: models
19
+ data_files: models.csv
20
+ - config_name: estimates
21
+ data_files: estimates.csv
22
+ ---
23
+
24
+ # llmledger
25
+
26
+ Every change to LiteLLM's model price catalogue since 6 September 2023: models
27
+ added and removed, prices changed, retirement dates set. Also one row per model,
28
+ and estimates of how likely a newly added model is to have its price changed,
29
+ get a retirement date, or be removed.
30
+
31
+ Data as of 13 October 2026, 01:26 UTC (LiteLLM commit `c64c1c8`). This is the
32
+ same data as llmledger version 1.0.
33
+
34
+ - Code, method, specification and tests: https://github.com/pouriaetab/llmledger
35
+ - Live results, rebuilt every day: https://pouriaetab.github.io/llmledger/
36
+ - Archived release: https://doi.org/10.5281/zenodo.23330454
37
+
38
+ > **Not financial or procurement advice.** The data describes edits to a public
39
+ > catalogue, not what any model costs you.
40
+
41
+ ## Files
42
+
43
+ ### `changes.csv`: one row per change (16,627 rows)
44
+
45
+ | Column | Meaning |
46
+ |:--|:--|
47
+ | `when` | Time of the LiteLLM commit, UTC |
48
+ | `sha` | LiteLLM commit |
49
+ | `model` | Model name as LiteLLM lists it |
50
+ | `kind` | `start` (in the first version), `added`, `removed`, `price` (a field whose name contains cost or pricing changed), `deprecation` (the retirement date changed) |
51
+ | `field` | The field that changed, for `price` and `deprecation` rows |
52
+ | `old`, `new` | Value before and after, as JSON (`null` when absent) |
53
+ | `flag` | `bulk` (removed on a day with 50 or more removals), `unit` (a price made 100 times larger or smaller), `reverted` (undone within 30 days), or empty |
54
+
55
+ ### `models.csv`: one row per model (5,275 rows)
56
+
57
+ Model name, a grouped maker and name from naming rules (for example,
58
+ `eu.anthropic.claude-3-5-sonnet-20241022-v2:0` becomes anthropic /
59
+ claude-3.5-sonnet), host, mode, status, dates added and removed, retirement
60
+ date, input and output price in US dollars per million tokens, number of price
61
+ changes (leaving out flagged ones), first exit, and age in days for models still
62
+ listed with no exit.
63
+
64
+ ### `estimates.csv`: competing risks estimates (80 rows)
65
+
66
+ Aalen-Johansen cumulative incidence of each first exit (price change,
67
+ retirement, removal, none) at 30, 90, 180, 365 and 730 days after a model was
68
+ added, with 95% bootstrap intervals (1,000 resamples) and the number of models
69
+ still followed. Four definitions: `main`, `headline` (input and output token
70
+ prices only), `no_first_week`, and `real_only` (leaving out flagged changes).
71
+
72
+ ## How it was built
73
+
74
+ Every version of `model_prices_and_context_window.json` on LiteLLM's `main`
75
+ branch was read in order with git and compared with the one before. The full
76
+ method, definitions and known limitations are in the
77
+ [specification](https://github.com/pouriaetab/llmledger/blob/main/docs/SPEC.md).
78
+
79
+ ## Citation
80
+
81
+ Emami Tabrizi, P. (2026). llmledger: what happens to an LLM after it joins a
82
+ price catalogue (version 1.0.0). Zenodo. https://doi.org/10.5281/zenodo.23330454
83
+
84
+ ## Licence and source
85
+
86
+ The data is derived from
87
+ [BerriAI/litellm](https://github.com/BerriAI/litellm), which is available under
88
+ the MIT licence below, and is shared under the same licence. llmledger is
89
+ independent and not affiliated with or endorsed by BerriAI.
90
+
91
+ ```
92
+ MIT License
93
+
94
+ Copyright (c) 2023 Berri AI
95
+
96
+ Permission is hereby granted, free of charge, to any person obtaining a copy
97
+ of this software and associated documentation files (the "Software"), to deal
98
+ in the Software without restriction, including without limitation the rights
99
+ to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
100
+ copies of the Software, and to permit persons to whom the Software is
101
+ furnished to do so, subject to the following conditions:
102
+
103
+ The above copyright notice and this permission notice shall be included in all
104
+ copies or substantial portions of the Software.
105
+
106
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
107
+ IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
108
+ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
109
+ AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
110
+ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
111
+ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
112
+ SOFTWARE.
113
+ ```
changes.csv ADDED
The diff for this file is too large to render. See raw diff
 
estimates.csv ADDED
@@ -0,0 +1,81 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ definition,days,outcome,estimate,ci_low,ci_high,at_risk
2
+ main,30,price,0.059072,0.05261,0.06606,3682
3
+ main,30,retirement,0.035134,0.03019,0.040379,3682
4
+ main,30,removed,0.10185,0.093864,0.110636,3682
5
+ main,30,none,0.803944,0.792771,0.814776,3682
6
+ main,90,price,0.085579,0.077579,0.093629,2717
7
+ main,90,retirement,0.05186,0.045302,0.058636,2717
8
+ main,90,removed,0.132793,0.123809,0.142673,2717
9
+ main,90,none,0.729768,0.717657,0.741825,2717
10
+ main,180,price,0.108513,0.098835,0.118261,2336
11
+ main,180,retirement,0.077049,0.069372,0.0849,2336
12
+ main,180,removed,0.141605,0.132812,0.151232,2336
13
+ main,180,none,0.672832,0.658483,0.686147,2336
14
+ main,365,price,0.146828,0.13538,0.158653,1214
15
+ main,365,retirement,0.163993,0.15198,0.177241,1214
16
+ main,365,removed,0.163389,0.15281,0.174451,1214
17
+ main,365,none,0.525791,0.50923,0.542899,1214
18
+ main,730,price,0.182766,0.169272,0.19692,379
19
+ main,730,retirement,0.295194,0.276343,0.313827,379
20
+ main,730,removed,0.23362,0.217222,0.249913,379
21
+ main,730,none,0.288421,0.267706,0.312241,379
22
+ headline,30,price,0.051013,0.044901,0.057234,3699
23
+ headline,30,retirement,0.035171,0.030217,0.040408,3699
24
+ headline,30,removed,0.10306,0.094863,0.112099,3699
25
+ headline,30,none,0.810757,0.799502,0.821612,3699
26
+ headline,90,price,0.075741,0.068334,0.083188,2736
27
+ headline,90,retirement,0.051962,0.045401,0.058737,2736
28
+ headline,90,removed,0.134349,0.125313,0.14427,2736
29
+ headline,90,none,0.737949,0.725446,0.749954,2736
30
+ headline,180,price,0.095715,0.087056,0.105212,2353
31
+ headline,180,retirement,0.079214,0.071308,0.087116,2353
32
+ headline,180,removed,0.143205,0.134326,0.152901,2353
33
+ headline,180,none,0.681866,0.668453,0.694705,2353
34
+ headline,365,price,0.125301,0.114968,0.136559,1240
35
+ headline,365,retirement,0.168095,0.155862,0.181383,1240
36
+ headline,365,removed,0.166198,0.155605,0.177206,1240
37
+ headline,365,none,0.540405,0.52353,0.55738,1240
38
+ headline,730,price,0.158343,0.145157,0.17206,381
39
+ headline,730,retirement,0.307056,0.287782,0.325965,381
40
+ headline,730,removed,0.238387,0.22193,0.254907,381
41
+ headline,730,none,0.296214,0.275058,0.319636,381
42
+ no_first_week,30,price,0.029108,0.024243,0.034213,3817
43
+ no_first_week,30,retirement,0.035316,0.030389,0.040595,3817
44
+ no_first_week,30,removed,0.10469,0.096659,0.113617,3817
45
+ no_first_week,30,none,0.830886,0.820249,0.841357,3817
46
+ no_first_week,90,price,0.05595,0.049452,0.062893,2840
47
+ no_first_week,90,retirement,0.053218,0.046531,0.060081,2840
48
+ no_first_week,90,removed,0.135501,0.126759,0.145289,2840
49
+ no_first_week,90,none,0.75533,0.74263,0.767146,2840
50
+ no_first_week,180,price,0.079466,0.071263,0.088598,2444
51
+ no_first_week,180,retirement,0.078971,0.071354,0.087069,2444
52
+ no_first_week,180,removed,0.144228,0.135467,0.154026,2444
53
+ no_first_week,180,none,0.697334,0.683507,0.71083,2444
54
+ no_first_week,365,price,0.120479,0.110204,0.131301,1224
55
+ no_first_week,365,retirement,0.167625,0.155399,0.180727,1224
56
+ no_first_week,365,removed,0.166628,0.156466,0.177887,1224
57
+ no_first_week,365,none,0.545268,0.528133,0.562385,1224
58
+ no_first_week,730,price,0.157453,0.144336,0.170861,379
59
+ no_first_week,730,retirement,0.304983,0.285422,0.324293,379
60
+ no_first_week,730,removed,0.240236,0.22384,0.256998,379
61
+ no_first_week,730,none,0.297328,0.276258,0.321523,379
62
+ real_only,30,price,0.03491,0.029856,0.040486,3802
63
+ real_only,30,retirement,0.035993,0.030883,0.041335,3802
64
+ real_only,30,removed,0.029767,0.025236,0.034892,3802
65
+ real_only,30,none,0.89933,0.890842,0.907598,3802
66
+ real_only,90,price,0.062317,0.055,0.070238,2830
67
+ real_only,90,retirement,0.054863,0.047962,0.061913,2830
68
+ real_only,90,removed,0.037646,0.032522,0.043193,2830
69
+ real_only,90,none,0.845174,0.833861,0.855877,2830
70
+ real_only,180,price,0.086252,0.077444,0.095841,2451
71
+ real_only,180,retirement,0.082863,0.074523,0.091085,2451
72
+ real_only,180,removed,0.045572,0.039678,0.052004,2451
73
+ real_only,180,none,0.785314,0.771573,0.797463,2451
74
+ real_only,365,price,0.123623,0.11299,0.135509,1246
75
+ real_only,365,retirement,0.186778,0.172303,0.201409,1246
76
+ real_only,365,removed,0.057346,0.050291,0.064892,1246
77
+ real_only,365,none,0.632254,0.615187,0.649981,1246
78
+ real_only,730,price,0.16798,0.152916,0.183095,382
79
+ real_only,730,retirement,0.349766,0.32749,0.372938,382
80
+ real_only,730,removed,0.093993,0.082415,0.107088,382
81
+ real_only,730,none,0.388262,0.362805,0.413436,382
models.csv ADDED
The diff for this file is too large to render. See raw diff