Models

101 models, one API key. Compare text, image, video and audio models by what they do and what they cost.

CompareModelPrecisionCapabilities
1M$0.27$0.8137tok/s0.69s
100%
1M$13.50$67.50$1.3525tok/s1.52s
100%
1M$1.013
$2.026
$5.063
$10.126
35tok/s1.95s
99.4%
1M$13.50$67.5013tok/s3.24s
77.0%
$0.006075 / min5.8×rt1.89s
100%
$0.00675 / min2.9×rt3.54s
66.7%
1M$0.203$0.63544tok/s2.60s
83.1%
1M$0.101
$0.203
$0.338
$0.675
30tok/s2.66s
95.6%
1M$0.297$0.89159tok/s1.36s
95.2%
1M$1.89$5.9432tok/s2.54s
98.6%
1M$0.567$4.0546tok/s1.88s
94.4%
1M$1.013
$2.026
$5.063
$10.126
$0.05062587tok/s2.98s
73.9%
500K$2.70$8.1056tok/s5.27s
91.7%
131K$0.405$1.6276tok/s3.79s
87.4%
1M$1.688$5.738130tok/s3.66s
89.3%
1M$2.2275$6.684$0.278127tok/s3.06s
97.9%
1M$0.189$0.37825tok/s3.56s
92.5%
1M$0.0498$0.2164$0.0081118tok/s2.33s
84.8%
1M$6.75$33.75$0.67521tok/s1.80s
79.6%
1M$1.013$5.063$0.1012526tok/s2.57s
92.5%
1M$0.405$3.375$0.040553tok/s1.13s
98.1%
1M$1.542$5.243185tok/s1.93s
89.3%
1M$4.1243$20.6213$0.40527tok/s2.81s
72.6%
1M$2.70
$5.40
$13.50
$27.00
$0.337555tok/s3.52s
91.2%
1M$2.70
$5.40
$13.50
$27.00
$0.337529tok/s1.53s
93.2%
1M$2.70$16.20$0.2774tok/s3.00s
78.9%
1M$2.70$16.20$0.2737tok/s1.24s
90.5%
1M$0.27$1.62$0.02775tok/s3.08s
85.8%
1M$0.27$1.62$0.02738tok/s1.19s
90.5%
500K$2.73$8.18949tok/s2.45s
91.7%
262K$0.108$0.44671tok/s4.36s
88.2%
1M$2.70$13.50$0.2717tok/s4.62s
89.6%
1M$1.101$3.76$0.29835fp4 · fp886tok/s2.34s
98.5%
262K$1.009$4.77437tok/s2.33s
98.4%
1M$13.50$67.5016tok/s3.17s
92.3%
1M$0.675$3.375$0.135fp4 · fp845tok/s2.08s
98.4%
1M$0.4431$1.773$0.086455tok/s17.58s
62.1%
1M$0.3985$1.594$0.0756fp4 · fp834tok/s1.93s
98.4%
256K$0.27$1.553$0.054fp852tok/s2.88s
98.0%
1M$2.304$6.909$0.1687563tok/s13.89s
72.5%
256K$1.389$2.777$0.2793tok/s3.07s
91.7%
1M$2.025$12.15$0.2025120tok/s2.30s
87.4%
1M$1.92$3.838$0.2785tok/s2.27s
92.8%
1M$0.1389$0.2778$0.024327tok/s1.43s
99.9%
$0.072 / image41s/img41.16s
40.9%
262K$0.7856$3.66739tok/s1.06s
96.5%
1M$6.75$33.75$0.67525tok/s1.70s
91.2%
203K$1.89$5.94$0.27675fp4 · fp829tok/s11.69s
98.4%
1M$0.4911$2.94750tok/s8.23s
74.7%
128K$0.0763$0.218$0.135fp4 · fp88tok/s6.72s
99.4%
2M$1.767$3.534329tok/s7.07s
93.3%
2M$1.763$3.52839tok/s1.10s
92.3%
205K$0.405$1.6257tok/s4.95s
99.7%
$0.061 / image
1M$2.70$16.20$0.2795tok/s4.83s
94.9%
$0.108 / image
1M$4.05$20.25$0.40531tok/s1.92s
93.1%
$0.054 / image
$0.3375 / image
197K$0.405$1.6272tok/s3.59s
99.2%
203K$0.081$0.54$0.0135fp852tok/s4.94s
94.3%
1M$0.675$4.05$0.067540tok/s1.43s
98.9%
$0.002592 / min
160K$0.351$0.513fp411tok/s1.24s
79.9%
$0.0405 / image13s/img12.92s
200K$1.35$6.75$0.13539tok/s1.46s
93.7%
2K$0.0027$0.001211tok/s0.33s
100%
$0.053 / image
128K$0.0494$0.240636tok/s3.40s
91.4%
131K$0.1945$1.272$0.03375fp838tok/s3.34s
97.2%
131K$0.334$1.519$0.135fp4 · fp849tok/s1.15s
99.2%
33K$0.0135$0.00253tok/s1.59s
99.1%
33K$0.0135$0.001152tok/s0.35s
100%
33K$0.027$0.001139tok/s0.35s
100%
41K$0.135$0.00
$0.054 / image
33K$0.108$0.37832tok/s18.48s
83.0%
1M$0.27$1.08fp834tok/s1.07s
91.8%
$0.07 / 1K char39ch/s3.21s
87.0%
$0.04 / 1K char41ch/s3.10s
97.8%
$0.081 / image
1M$0.405$3.375$0.040540tok/s1.12s
98.6%
200K$4.05$20.2527tok/s1.66s
92.3%
127K$1.35$1.3527tok/s1.64s
95.8%
$0.005 / image30s/img30.07s
100%
64K$0.7425$2.95773tok/s1.58s
100%
128K$0.135$0.43216tok/s0.88s
88.9%
$0.2025 / 1K char111ch/s1.14s
$0.0009 / min8.8×rt1.15s
100%
131K$1.35$1.35fp817tok/s0.89s
100%
131K$0.945$0.945fp825tok/s0.85s
96.0%
$0.2025 / 1K char109ch/s1.15s
$0.00405 / min5.1×rt1.96s
99.4%
1K$0.0135$0.00460tok/s1.11s
100%
8K$0.0135$0.001318tok/s0.40s
100%
1K$0.0135$0.001381tok/s0.29s
100%
1K$0.00675$0.001500tok/s0.27s
100%
$0.405 / 1K char56ch/s2.26s
100%
1K$0.00675$0.00449tok/s0.90s
100%
1K$0.00675$0.0010tok/s12.58s
100%
1K$0.00675$0.00259tok/s0.99s
100%

Token models priced per 1M tokens. The cached column is what a repeated prompt prefix costs on that model — a rate of its own, read from the same source as the input rate beside it, not a fixed fraction of it. It runs from a tenth of the input rate to the full input rate depending on the model, and a dash means nobody publishes one, which is a different statement from “free”. Image, video, and audio models priced per unit (image, second, video, minute, song, call, or 1K characters). Throughput and latency are a probe with an identical input when Kyma has that sample. If a row is labeled “from traffic, 7d”, that number is production traffic instead — real prompts, not comparable across models. Image throughput is seconds per succeeded job (30-day median). Video, music and realtime stay empty until a job exists. An empty throughput, latency or uptime cell means Kyma does not yet have a number for that cell. Uptime is a percentage only after 20 observations, over 30 days, probes and real requests together, including Kyma's own accounts. Rankings demand is the same total. Precision names the numeric formats a model may be served at that are BELOW the one its creator released it in, and it is filled on the 14 models where that is true. A low format is not a downgrade when it is how the model shipped, so this is a comparison against the creator’s own model card rather than a list of small numbers. An empty precision cell means there is nothing to flag: either no route reports going under the release, or nobody could source what the release was. Those two are different answers and the model’s own page prints which one it is.

Start building with Kyma API

$0.50 signup credit for eligible free-tier models after email verification.

Create account