Models

126 models, one API key. Open-source text, code, image, and video — pay per token, image, or second.

CompareModelPrecisionCapabilities
1M$1.013
$2.026
$5.063
$10.126
35tok/s1.91s
99.2%
1M$13.50$67.5013tok/s3.28s
71.4%
$0.006075 / min5.9×rt1.87s
no data yet
$0.00675 / min2.9×rt3.47s
no data yet
1M$0.203$0.63546tok/s2.41s
73.3%
1M$0.101
$0.203
$0.338
$0.675
30tok/s2.64s
89.1%
1M$0.297$0.89192tok/s1.03s
92.9%
1M$1.89$5.9432tok/s2.38s
96.9%
1M$0.567$4.0546tok/s1.72s
93.1%
1M$1.013
$2.026
$5.063
$10.126
$0.05062590tok/s2.95s
73.9%
500K$2.70$8.1057tok/s4.92s
93.5%
131K$0.405$1.6280tok/s3.69s
83.9%
1M$1.688$5.738133tok/s3.60s
80.3%
1M$2.2275$6.684$0.278128tok/s2.98s
97.2%
1M$0.189$0.37824tok/s4.03s
90.3%
1M$0.0498$0.2164$0.0081110tok/s2.44s
86.9%
1M$13.50$67.50$1.3534tok/s1.38s
93.2%
1M$6.75$33.75$0.67520tok/s1.84s
87.6%
1M$1.013$5.063$0.1012530tok/s2.52s
98.1%
1M$0.405$3.375$0.040554tok/s1.10s
98.0%
1M$1.542$5.243184tok/s1.93s
92.2%
1M$4.05$20.25$0.40529tok/s2.62s
76.0%
1M$2.70
$5.40
$13.50
$27.00
$0.337554tok/s3.69s
93.9%
1M$2.70
$5.40
$13.50
$27.00
$0.337529tok/s1.55s
97.3%
1M$2.70$16.20$0.2776tok/s2.95s
84.2%
1M$2.70$16.20$0.2738tok/s1.22s
93.6%
1M$0.27$1.62$0.02776tok/s3.01s
90.5%
1M$0.27$1.62$0.02739tok/s1.17s
92.0%
500K$2.73$8.18949tok/s2.46s
93.9%
262K$0.108$0.44674tok/s3.91s
84.9%
1M$2.70$13.50$0.2718tok/s4.38s
95.3%
1M$1.101$3.76$0.29835fp4 · fp882tok/s2.74s
98.5%
262K$1.009$4.77442tok/s2.06s
98.9%
1M$13.50$67.5015tok/s3.46s
93.7%
$0.06345 / minno data yetno data yet
no data yet
1M$0.675$3.375$0.135fp4 · fp846tok/s1.69s
98.8%
1M$0.4431$1.773$0.086454tok/s17.02s
60.8%
1M$0.3985$1.594$0.0756fp4 · fp834tok/s1.93s
98.2%
256K$0.27$1.553$0.054fp851tok/s2.92s
98.4%
1M$2.304$6.909$0.1687559tok/s15.10s
74.5%
256K$1.389$2.777$0.2799tok/s2.83s
94.0%
1M$2.025$12.15$0.2025123tok/s2.25s
87.7%
$0.0459 / minno data yetno data yet
no data yet
1M$1.92$3.838$0.2786tok/s2.22s
94.8%
1M$0.6901$1.38$0.13331tok/s3.48s
99.0%
1M$0.1389$0.2778$0.024328tok/s1.42s
99.9%
$0.072 / imageno data yetno data yet
38.4%
262K$0.7856$3.66739tok/s1.06s
96.7%
1M$6.75$33.75$0.67524tok/s1.71s
93.6%
$0.40959 / secno data yetno data yet
no data yet
$0.326565 / secno data yetno data yet
no data yet
$0.20 / songno data yetno data yet
no data yet
203K$1.89$5.94$0.27675fp4 · fp829tok/s11.78s
98.7%
1M$0.4911$2.94750tok/s8.24s
74.6%
128K$0.0763$0.218$0.135fp4 · fp88tok/s6.63s
99.4%
2M$1.767$3.534314tok/s7.44s
95.2%
2M$1.763$3.52839tok/s1.08s
94.5%
$0.03888 / minno data yetno data yet
no data yet
205K$0.405$1.6257tok/s4.89s
91.6%
$0.061 / imageno data yetno data yet
no data yet
1M$2.70$16.20$0.2795tok/s4.83s
94.8%
$0.108 / imageno data yetno data yet
no data yet
1M$4.05$20.25$0.40531tok/s1.92s
94.8%
$0.054 / imageno data yetno data yet
no data yet
$0.3375 / imageno data yetno data yet
no data yet
$0.405 / imageno data yetno data yet
no data yet
197K$0.3826$1.34673tok/s3.65s
100%
$0.1512 / secno data yetno data yet
no data yet
$0.2268 / secno data yetno data yet
no data yet
262K$0.6075$3.03835tok/s5.45s
91.9%
203K$0.081$0.54$0.0135fp861tok/s4.55s
96.3%
1M$0.675$4.05$0.067540tok/s1.40s
98.9%
$0.002592 / minno data yetno data yet
no data yet
$0.03888 / minno data yetno data yet
no data yet
160K$0.351$0.513fp411tok/s1.21s
80.7%
$0.0405 / imageno data yetno data yet
no data yet
$0.20 / songno data yetno data yet
no data yet
200K$1.35$6.75$0.13540tok/s1.42s
93.9%
$0.0945 / secno data yetno data yet
no data yet
2K$0.0027$0.001239tok/s0.32s
100%
$0.053 / imageno data yetno data yet
no data yet
$0.135 / secno data yetno data yet
no data yet
128K$0.0494$0.240634tok/s3.61s
93.1%
$0.135 / secno data yetno data yet
no data yet
131K$0.1945$1.272$0.03375fp837tok/s3.15s
97.8%
131K$0.334$1.519$0.135fp4 · fp849tok/s1.14s
99.1%
$2.00 / callno data yetno data yet
no data yet
$0.78 / videono data yetno data yet
no data yet
$0.42 / videono data yetno data yet
no data yet
$0.14 / videono data yetno data yet
no data yet
$0.405 / 1K char31ch/s4.05s
no data yet
33K$0.0135$0.00260tok/s1.54s
99.7%
33K$0.0135$0.001156tok/s0.35s
100%
33K$0.027$0.001181tok/s0.34s
100%
41K$0.135$0.00no data yetno data yet
no data yet
$0.054 / imageno data yetno data yet
no data yet
$0.54 / secno data yetno data yet
no data yet
33K$0.108$0.37839tok/s13.42s
90.5%
1M$0.27$1.08fp835tok/s1.08s
94.0%
$2.00 / callno data yetno data yet
no data yet
$0.07 / 1K char40ch/s3.18s
88.6%
$0.04 / 1K char41ch/s3.10s
98.8%
$0.081 / imageno data yetno data yet
no data yet
1M$0.405$3.375$0.040541tok/s1.09s
96.5%
200K$4.05$20.2526tok/s1.66s
93.8%
127K$1.35$1.3526tok/s1.68s
95.9%
$0.005 / imageno data yetno data yet
45.7%
64K$0.7425$2.95775tok/s1.63s
99.8%
128K$0.135$0.43216tok/s0.89s
91.0%
$0.081 / imageno data yetno data yet
no data yet
$0.054 / imageno data yetno data yet
no data yet
$0.2025 / 1K char111ch/s1.13s
no data yet
$0.0009 / min9.0×rt1.13s
100%
131K$1.35$1.35fp817tok/s0.94s
100%
131K$0.945$0.945fp825tok/s0.84s
96.9%
$0.2025 / 1K char110ch/s1.15s
no data yet
$0.027 / callno data yetno data yet
no data yet
$0.00405 / min5.1×rt1.95s
99.0%
1K$0.0135$0.00456tok/s1.12s
100%
8K$0.0135$0.001332tok/s0.39s
98.5%
1K$0.0135$0.001449tok/s0.28s
100%
1K$0.00675$0.001534tok/s0.26s
100%
$0.405 / 1K char56ch/s2.24s
no data yet
1K$0.00675$0.00448tok/s0.90s
100%
1K$0.00675$0.009tok/s14.29s
100%
1K$0.00675$0.00262tok/s0.98s
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 measured by Kyma every six hours with an identical input sent to every model through this API — so the numbers compare. For chat and embedding models throughput is tokens per second; for transcription models it is the realtime factor (×rt: seconds of audio transcribed per second of wall time); for text-to-speech it is characters synthesised per second (ch/s) — each off one fixed input. Image, video, music and realtime models are not probed; their reliability comes from real jobs, so those two columns stay empty by design. Uptime is every observation over 30 days, probes and real requests together. 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 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 any model

$0.50 free credit on the free tier. OpenAI-compatible API.

Get Free API Key