The Known Good Updated 29 Jul 2026
Providers· Inference host· 68 tracked models

Novita — pricing & performance

Compare models All providers Data sources
Models served
68
of 947 tracked models
Mean output speed
38
tokens per second, averaged over
66 measured endpoints
Mean latency
1.7s
time to first token, averaged over
66 measured endpoints
Cheapest blended
$0.000
per 1M tokens at a 3:1
input:output blend
Output speed on Novita

Median output tokens per second for the 66 of 68 models on Novita with published throughput telemetry from OpenRouter. The full table below lists all 68, measured or not. Higher is better. We do not run these measurements ourselves.

⇩
126
117
99
92
83
77
76
73
68
67
66
64
The Known Good
68 models on Novita
Full leaderboard →
Model Creator Input $/M Output $/M Blended Speed TTFT Context
Baidu: ERNIE 4.5 VL 424B A47B Baidu $0.42 $1.25 $0.63 7 1.2s 123k
DeepSeek: DeepSeek V3 DeepSeek $0.40 $1.30 $0.62 20 1.4s 64k
DeepSeek: DeepSeek V3 0324 DeepSeek $0.27 $1.12 $0.48 29 1.5s 164k
DeepSeek: DeepSeek V3.1 DeepSeek $0.27 $1.00 $0.45 17 1.7s 131k
DeepSeek: DeepSeek V3.1 Terminus DeepSeek $0.27 $1.00 $0.45 25 1.6s 131k
DeepSeek: DeepSeek V3.2 DeepSeek $0.27 $0.40 $0.30 22 1.1s 164k
DeepSeek: DeepSeek V3.2 Exp DeepSeek $0.27 $0.41 $0.30 20 1.5s 164k
DeepSeek: DeepSeek V4 Flash DeepSeek $0.14 $0.28 $0.17 58 1.4s 1.05M
DeepSeek: DeepSeek V4 Pro DeepSeek $1.17 $2.34 $1.46 62 1.6s 1.05M
DeepSeek: R1 DeepSeek $0.70 $2.50 $1.15 32 0.9s 64k
DeepSeek: R1 0528 DeepSeek $0.70 $2.50 $1.15 — — 164k
DeepSeek: R1 Distill Llama 70B DeepSeek $0.80 $0.80 $0.80 32 0.8s 8k
Google: Gemma 3 27B Google $0.12 $0.20 $0.14 17 2.0s 98k
Google: Gemma 4 26B A4B Google $0.13 $0.40 $0.20 15 1.7s 262k
Google: Gemma 4 31B Google $0.14 $0.40 $0.20 19 1.2s 262k
inclusionAI: Ling-2.6-1T inclusionai $0.075 $0.62 $0.21 3 1.1s 262k
inclusionAI: Ling-2.6-flash inclusionai $0.010 $0.030 $0.015 77 0.6s 262k
inclusionAI: Ring-2.6-1T inclusionai $0.075 $0.62 $0.21 76 1.2s 262k
Ling-3.0-flash (free) inclusionai $0.000 $0.000 $0.000 92 1.9s 262k
Meta: Llama 3.1 8B Instruct Meta $0.020 $0.050 $0.028 83 0.5s 16k
Meta: Llama 3.3 70B Instruct Meta $0.14 $0.40 $0.20 31 0.6s 6k
Meta: Llama 4 Maverick Meta $0.27 $0.85 $0.41 20 0.5s 1.05M
Meta: Llama 4 Scout Meta $0.18 $0.59 $0.28 22 0.6s 131k
MiniMax: MiniMax M1 MiniMax $0.55 $2.20 $0.96 29 1.6s 1.00M
MiniMax: MiniMax M2 MiniMax $0.30 $1.20 $0.53 68 1.1s 205k
MiniMax: MiniMax M2.1 MiniMax $0.30 $1.20 $0.53 41 1.9s 205k
MiniMax: MiniMax M2.5 MiniMax $0.30 $1.20 $0.53 10 1.2s 205k
MiniMax: MiniMax M2.7 MiniMax $0.27 $1.08 $0.47 29 2.2s 205k
MiniMax: MiniMax M3 MiniMax $0.30 $1.20 $0.53 42 2.2s 1.00M
Mistral: Mistral Nemo Mistral $0.040 $0.17 $0.072 13 1.6s 60k
MoonshotAI: Kimi K2 0711 Moonshot AI $0.57 $2.30 $1.00 29 0.7s 131k
MoonshotAI: Kimi K2 0905 Moonshot AI $0.60 $2.50 $1.07 16 0.8s 262k
MoonshotAI: Kimi K2 Thinking Moonshot AI $0.60 $2.50 $1.07 33 1.1s 262k
MoonshotAI: Kimi K2.5 Moonshot AI $0.57 $2.85 $1.14 30 2.4s 262k
MoonshotAI: Kimi K2.6 Moonshot AI $0.80 $3.40 $1.45 31 1.5s 262k
MoonshotAI: Kimi K2.7 Code Moonshot AI $0.91 $3.84 $1.64 2 1.1s 262k
NVIDIA: Nemotron 3 Nano 30B A3B NVIDIA $0.050 $0.20 $0.087 126 0.7s 262k
OpenAI: gpt-oss-120b OpenAI $0.050 $0.25 $0.10 64 0.6s 131k
OpenAI: gpt-oss-20b OpenAI $0.040 $0.15 $0.068 117 3.9s 131k
Qwen2.5 72B Instruct Alibaba $0.38 $0.40 $0.39 — — 32k
Qwen: Qwen3 235B A22B Instruct 2507 Alibaba $0.090 $0.58 $0.21 26 0.7s 131k
Qwen: Qwen3 235B A22B Thinking 2507 Alibaba $0.30 $3.00 $0.97 23 1.3s 131k
Qwen: Qwen3 Coder 30B A3B Instruct Alibaba $0.070 $0.27 $0.12 60 0.5s 160k
Qwen: Qwen3 Coder 480B A35B Alibaba $0.38 $1.55 $0.67 36 1.3s 262k
Qwen: Qwen3 Coder Next Alibaba $0.20 $1.50 $0.53 32 1.6s 262k
Qwen: Qwen3 Next 80B A3B Instruct Alibaba $0.15 $1.50 $0.49 67 0.9s 131k
Qwen: Qwen3 VL 235B A22B Instruct Alibaba $0.30 $1.50 $0.60 23 1.2s 131k
Qwen: Qwen3 VL 235B A22B Thinking Alibaba $0.98 $3.95 $1.72 40 0.8s 131k
Qwen: Qwen3 VL 30B A3B Instruct Alibaba $0.20 $0.70 $0.33 5 0.8s 131k
Qwen: Qwen3.5 397B A17B Alibaba $0.60 $3.60 $1.35 58 1.5s 262k
Qwen: Qwen3.5-122B-A10B Alibaba $0.40 $3.20 $1.10 73 1.0s 262k
Qwen: Qwen3.5-27B Alibaba $0.30 $2.40 $0.82 14 1.0s 262k
Sao10K: Llama 3 8B Lunaris sao10k $0.050 $0.050 $0.050 66 0.3s 8k
Sao10K: Llama 3.1 Euryale 70B v2.2 sao10k $1.45 $1.45 $1.45 39 0.6s 8k
StepFun: Step 3.7 Flash stepfun $0.20 $1.15 $0.44 99 3.8s 262k
Tencent: Hy3 Tencent $0.14 $0.58 $0.25 49 3.1s 262k
WizardLM-2 8x22B Microsoft $0.62 $0.62 $0.62 17 0.8s 66k
Xiaomi: MiMo-V2.5 Xiaomi $0.17 $0.34 $0.21 49 3.2s 1.05M
Xiaomi: MiMo-V2.5-Pro Xiaomi $0.48 $0.96 $0.60 39 2.2s 1.05M
Z.ai: GLM 4.5 Air Z.ai $0.13 $0.85 $0.31 31 0.7s 131k
Z.ai: GLM 4.5V Z.ai $0.60 $1.80 $0.90 41 5.7s 66k
Z.ai: GLM 4.6 Z.ai $0.55 $2.20 $0.96 36 1.5s 205k
Z.ai: GLM 4.6V Z.ai $0.30 $0.90 $0.45 10 1.8s 131k
Z.ai: GLM 4.7 Z.ai $0.54 $1.98 $0.90 21 2.3s 205k
Z.ai: GLM 4.7 Flash Z.ai $0.070 $0.40 $0.15 6 16.4s 200k
Z.ai: GLM 5 Z.ai $1.00 $3.20 $1.55 28 2.2s 203k
Z.ai: GLM 5.1 Z.ai $1.38 $4.40 $2.13 51 1.8s 205k
Z.ai: GLM 5.2 Z.ai $0.76 $2.38 $1.16 27 2.6s 1.05M
The Known Good
Source. Prices, endpoint availability and throughput for Novita are ingested from OpenRouter; model metadata is enriched from models.dev. A blank cell means the figure was not published for that endpoint — it never means zero. Blended price is (3 × input + output) ÷ 4, the same 3:1 blend used everywhere on this site; see methodology. Quality scores are not provider-specific: a model scores the same wherever it is hosted, so quality lives on the model pages.