API host · catalogue checked 27 Oct 2026

Novita: models, prices and uptime

Novita serves 19 of the 78 models we benchmark through OpenRouter. Prices are OpenRouter's list prices per million tokens. We did not measure speed per host.

Models served19of 78 we test

Cheapest host for2of 19 shared models

Median premium+54%over the cheapest host

Uptime100%last 30 min, average

Most accurate models on Novita

Verification score of each model this host serves

1DeepSeek V4.1 Flash 1017DeepSeek73.8%
2DeepSeek V4 Pro 0813DeepSeek71.2%
3MiMo-V2.6-ProXiaomi68.8%
4GLM 5.3 FlashZ.ai67.5%
5GLM 5.3Z.ai67.5%
6MiniMax M3.1 FlashMiniMax66.2%
7Kimi K2.7 CodeMoonshotAI65.0%
8Qwen3.8 27BQwen63.8%
9MiMo-V2.6-FlashXiaomi62.5%
10Hy3Tencent58.8%
11gpt-oss-120bOpenAI58.8%
12Qwen3.8 2.4T A95BQwen57.5%
13MiniMax M3MiniMax57.5%
14Llama 3.3 70B InstructMeta42.5%
15Ling 3.1 FlashinclusionAI41.2%
16Qwen3.5 397B A17BQwen32.5%
17Gemma 4 31BGoogle30.0%
18Gemma 4 26B A4BGoogle15.0%

Models on Novita

Premium: this host's blended price (3 input : 1 output) against the cheapest host for the same model

Model	Score	Input / 1M	Output / 1M	vs cheapest host	Hosts	Context	Max output	Quantization	Uptime
DeepSeek V4.1 Flash 1017	73.8%	$0.195	$0.78	+201%	30	1M	393k	fp8	100%
DeepSeek V4 Pro 0813	71.2%	$0.99	$2.97	+608%	23	1M	393k	fp8	100%
MiMo-V2.6-Pro	68.8%	$0.435	$0.87	+18%	5	1M	131k	fp8	100%
GLM 5.3 Flash	67.5%	$0.084	$0.28	+183%	34	1M	131k	fp8	100%
GLM 5.3	67.5%	$0.63	$1.98	+140%	37	1M	131k	fp8	100%
MiniMax M3.1 Flash	66.2%	$0.15	$0.6	+5%	11	1M	524k	fp16	100%
Kimi K2.7 Code	65.0%	$0.912	$3.84	+38%	11	262k	236k	int4	100%
Qwen3.8 27B	63.8%	$0.42	$3	+526%	21	1M	131k	—	—
MiMo-V2.6-Flash	62.5%	$0.14	$0.28	+21%	10	1M	131k	fp8	100%
Hy3	58.8%	$0.14	$0.58	+73%	5	262k	236k	—	100%
gpt-oss-120b	58.8%	$0.05	$0.25	+67%	23	131k	33k	fp4	100%
Qwen3.8 2.4T A95B	57.5%	$2	$6	Cheapest	5	1M	131k	—	100%
MiniMax M3	57.5%	$0.3	$1.2	+35%	12	1M	131k	fp8	100%
Llama 3.3 70B Instruct	42.5%	$0.135	$0.4	+30%	9	12k	11k	bf16	100%
Ling 3.1 Flash	41.2%	$0.075	$0.225	Cheapest	2	262k	236k	—	100%
Qwen3.5 397B A17B	32.5%	$0.6	$3.6	+54%	10	262k	66k	—	—
Gemma 4 31B	30.0%	$0.14	$0.4	+28%	13	262k	131k	bf16	100%
Gemma 4 26B A4B	15.0%	$0.13	$0.4	+128%	15	262k	131k	bf16	100%
Qwen3.5-122B-A10B	—	$0.4	$3.2	+54%	4	262k	66k	bf16	100%

All API hosts →

Novita: frequently asked questions

Which AI models does Novita serve?: Of the models we benchmark, Novita serves 19 through OpenRouter: DeepSeek V4.1 Flash 1017, DeepSeek V4 Pro 0813, MiMo-V2.6-Pro, GLM 5.3 Flash, GLM 5.3, MiniMax M3.1 Flash, Kimi K2.7 Code, Qwen3.8 27B, MiMo-V2.6-Flash, Hy3, gpt-oss-120b, Qwen3.8 2.4T A95B, MiniMax M3, Llama 3.3 70B Instruct, Ling 3.1 Flash, Qwen3.5 397B A17B, Gemma 4 31B, Gemma 4 26B A4B, Qwen3.5-122B-A10B.
What is the most accurate model on Novita?: DeepSeek V4.1 Flash 1017 by DeepSeek, with a verification score of 73.8% (59 of 80 questions correct) in our tests.
Is Novita the cheapest host?: Novita has the lowest list price for 2 of the 19 models it serves that other hosts also serve (Qwen3.8 2.4T A95B, Ling 3.1 Flash). Its median markup over the cheapest host is +54%.
How reliable is Novita?: Averaged over the models it serves, Novita reported 100% uptime in the 30 minutes before OpenRouter's catalogue was read for our run.

Model

Score

Input / 1M

Output / 1M

vs cheapest host

Hosts

Context

Max output

Quantization

Uptime

DeepSeek V4.1 Flash 1017

73.8%

$0.195

$0.78

+201%

393k

fp8

100%

DeepSeek V4 Pro 0813

71.2%

$0.99

$2.97

+608%

393k

fp8

100%

MiMo-V2.6-Pro

68.8%

$0.435

$0.87

+18%

131k

fp8

100%

GLM 5.3 Flash

67.5%

$0.084

$0.28

+183%

131k

fp8

100%

GLM 5.3

67.5%

$0.63

$1.98

+140%

131k

fp8

100%

MiniMax M3.1 Flash

66.2%

$0.15

$0.6

+5%

524k

fp16

100%

Kimi K2.7 Code

65.0%

$0.912

$3.84

+38%

262k

236k

int4

100%

Qwen3.8 27B

63.8%

$0.42

+526%

131k

—

MiMo-V2.6-Flash

62.5%

$0.14

$0.28

+21%

131k

fp8

100%

Hy3

58.8%

$0.14

$0.58

+73%

262k

236k

—

100%

gpt-oss-120b

58.8%

$0.05

$0.25

+67%

131k

33k

fp4

100%

Qwen3.8 2.4T A95B

57.5%

Cheapest

131k

—

100%

MiniMax M3

57.5%

$0.3

$1.2

+35%

131k

fp8

100%

Llama 3.3 70B Instruct

42.5%

$0.135

$0.4

+30%

12k

11k

bf16

100%

Ling 3.1 Flash

41.2%

$0.075

$0.225

Cheapest

262k

236k

—

100%

Qwen3.5 397B A17B

32.5%

$0.6

$3.6

+54%

262k

66k

—

Gemma 4 31B

30.0%

$0.14

$0.4

+28%

262k

131k

bf16

100%

Gemma 4 26B A4B

15.0%

$0.13

$0.4

+128%

262k

131k

bf16

100%

Qwen3.5-122B-A10B

—

$0.4

$3.2

+54%

262k

66k

bf16

100%

Novita: frequently asked questions

Which AI models does Novita serve?

Of the models we benchmark, Novita serves 19 through OpenRouter: DeepSeek V4.1 Flash 1017, DeepSeek V4 Pro 0813, MiMo-V2.6-Pro, GLM 5.3 Flash, GLM 5.3, MiniMax M3.1 Flash, Kimi K2.7 Code, Qwen3.8 27B, MiMo-V2.6-Flash, Hy3, gpt-oss-120b, Qwen3.8 2.4T A95B, MiniMax M3, Llama 3.3 70B Instruct, Ling 3.1 Flash, Qwen3.5 397B A17B, Gemma 4 31B, Gemma 4 26B A4B, Qwen3.5-122B-A10B.

What is the most accurate model on Novita?

DeepSeek V4.1 Flash 1017 by DeepSeek, with a verification score of 73.8% (59 of 80 questions correct) in our tests.

Is Novita the cheapest host?

Novita has the lowest list price for 2 of the 19 models it serves that other hosts also serve (Qwen3.8 2.4T A95B, Ling 3.1 Flash). Its median markup over the cheapest host is +54%.

How reliable is Novita?

Averaged over the models it serves, Novita reported 100% uptime in the 30 minutes before OpenRouter's catalogue was read for our run.