Novita AI vs Runcrate
API prices for the 32 models both providers list, side by side. Prices as of October 25, 2026.
Novita AI is cheaper on 7 of the 32 models we can price on both, Runcrate on 24, and 1 cost the same. Runcrate is 15% cheaper than Novita AI on the median shared model.
Novita AI is an aggregator: it routes requests to models hosted by other providers, so a price there may be another provider's rate passed through.
Runcrate is an aggregator: it routes requests to models hosted by other providers, so a price there may be another provider's rate passed through.
Each side is its lowest listed price for the model. LLMs are compared on input + output price per 1M tokens; image, video and audio models on their per-unit price. Resolution, tier and variant can change the price. Methodology
At a glance
| Novita AI | Runcrate | |
|---|---|---|
| Type | Aggregator | Aggregator |
| Models we track | 68 | 100 |
| Free tier | Yes | DeepSeek V4.1 Flash, GLM 5.3, GLM 5.3 Flash, gpt-oss 120B and Qwen3.8 Flash free with a GitHub account at least 30 days old (session cap plus weekly allowance, resets Monday) |
| OpenAI-compatible API | ||
| Function calling | — | |
| JSON mode | — | |
| Vision | ||
| Batch API | — | — |
| Fine-tuning | — | — |
| SOC 2 | — | — |
| HIPAA | — | — |
| GDPR | — | — |
Prices for models on both
LLM prices are input / output per 1M tokens. The cheaper side is highlighted.
LLMs (32 · Novita AI cheaper on 7, Runcrate on 24)
| Model | Novita AI | Runcrate | Difference |
|---|---|---|---|
| DeepSeek R1 0528 DeepSeek | $0.700 / $2.50 | $0.525 / $2.26 | Runcrate −13% |
| DeepSeek V4 Flash DeepSeek | $0.140 / $0.280 | $0.095 / $0.189 | Runcrate −33% |
| DeepSeek V4 Pro DeepSeek | $1.60 / $3.20 | $1.36 / $2.73 | Runcrate −15% |
| DeepSeek V4.1 Flash DeepSeek | $0.300 / $1.20 | $0.147 / $0.441 | Runcrate −61% |
| GLM 4.7 Zhipu AI | $0.600 / $2.20 | $0.336 / $1.47 | Runcrate −36% |
| GLM 5.2 Z.ai | $1.40 / $4.40 | $0.591 / $1.89 | Runcrate −57% |
| GLM 5.3 Z.ai | $1.40 / $4.40 | $0.591 / $2.63 | Runcrate −45% |
| GLM 5.3 Flash Z.ai | $0.150 / $0.500 | $0.158 / $0.525 | Novita AI −5% |
| GPT OSS 120B OpenAI | $0.050 / $0.250 | $0.039 / $0.178 | Runcrate −28% |
| Gemma 3 12B Google | $0.050 / $0.100 | $0.042 / $0.126 | Novita AI −11% |
| Gemma 3 27B Google | $0.119 / $0.200 | $0.084 / $0.168 | Runcrate −21% |
| Gemma 4 26B A4B Google | $0.130 / $0.400 | $0.073 / $0.357 | Runcrate −19% |
| Gemma 4 31B Google | $0.140 / $0.400 | $0.142 / $0.535 | Novita AI −20% |
| Hy4 Preview Tencent | $0.834 / $2.50 | $0.876 / $2.63 | Novita AI −5% |
| Kimi K2.6 Moonshot AI | $0.800 / $3.40 | $0.787 / $3.67 | Novita AI −6% |
| Kimi K3 Moonshot AI | $3.00 / $15.00 | $2.99 / $14.96 | Same |
| Llama 3.3 70B Meta | $0.135 / $0.400 | $0.105 / $0.336 | Runcrate −18% |
| MiMo V2.6 Flash Xiaomi | $0.140 / $0.280 | $0.096 / $0.191 | Runcrate −32% |
| MiMo V2.6 Pro Xiaomi | $0.435 / $0.870 | $0.406 / $0.822 | Runcrate −6% |
| MiniMax M3 MiniMax | $0.300 / $1.20 | $0.294 / $1.16 | Runcrate −3% |
| Qwen3 Coder 480B A35B Alibaba | $0.380 / $1.55 | $0.315 / $1.05 | Runcrate −29% |
| Qwen3 Next 80B A3B Instruct Alibaba | $0.150 / $1.50 | $0.076 / $0.924 | Runcrate −39% |
| Qwen3 VL 235B A22B Instruct Alibaba | $0.300 / $1.50 | $0.210 / $0.924 | Runcrate −37% |
| Qwen3 VL 30B A3B Instruct Alibaba | $0.200 / $0.700 | $0.158 / $0.630 | Runcrate −12% |
| Qwen3.5-27B Alibaba | $0.300 / $2.40 | $0.273 / $2.73 | Novita AI −10% |
| Qwen3.6 27B Alibaba | $0.600 / $3.60 | $0.336 / $3.36 | Runcrate −12% |
| Qwen3.6 35B A3B Alibaba | $0.248 / $1.49 | $0.105 / $0.998 | Runcrate −36% |
| Qwen3.7 Max Alibaba | $1.25 / $3.75 | $2.63 / $7.88 | Novita AI −52% |
| Qwen3.8 27B Alibaba | $0.420 / $3.00 | $0.158 / $1.97 | Runcrate −38% |
| Qwen3.8 Flash Alibaba | $0.150 / $0.470 | $0.119 / $0.401 | Runcrate −16% |
| Qwen3.8 Max Alibaba | $2.00 / $6.00 | $1.73 / $5.20 | Runcrate −13% |
| gpt-oss-20b OpenAI | $0.040 / $0.150 | $0.032 / $0.147 | Runcrate −6% |
Only on Novita AI (36)
Only on Runcrate (68)
Novita AI vs Runcrate FAQ
Is Novita AI cheaper than Runcrate?
Usually no. Novita AI is cheaper on 7 of the 32 models we can price on both, Runcrate on 24, and 1 cost the same. Runcrate is 15% cheaper than Novita AI on the median shared model.
Which has more models, Novita AI or Runcrate?
We track 68 models on Novita AI and 100 on Runcrate. 32 are on both, 36 only on Novita AI and 68 only on Runcrate.
Do Novita AI and Runcrate have a free tier?
Novita AI: yes. Runcrate: yes (DeepSeek V4.1 Flash, GLM 5.3, GLM 5.3 Flash, gpt-oss 120B and Qwen3.8 Flash free with a GitHub account at least 30 days old (session cap plus weekly allowance, resets Monday)).
Which models do both Novita AI and Runcrate offer?
32 models, including DeepSeek R1 0528, DeepSeek V4 Flash, DeepSeek V4 Pro, DeepSeek V4.1 Flash, GLM 4.7, GLM 5.2 and more. The full list with both prices is in the table above.