MagmaRouter vs Novita AI
Novita AI is a large inference platform with hundreds of open models, OpenAI and Anthropic compatible endpoints, batch discounts and GPU rental, often at low per-token prices. MagmaRouter is narrower: a handful of uncensored Xploded models built to drive coding agents, with hard spend caps on every key.
Facts checked October 2026 against Novita AI’s own site, docs and policies.
| MagmaRouter | Novita AI | |
|---|---|---|
| What it is | Hosted gateway for one uncensored, tool capable model line (Xploded). | A cloud platform with APIs for LLMs, image, video and audio generation, plus GPU instances and agent sandboxes. |
| Pricing | Pay as you go, $5 minimum top-up, no subscription required. | Per token, input and output priced separately and varying by model (for example Llama 3.1 8B at $0.02 in and $0.05 out per million). Batch inference is 50% off on supported models, some models have cheaper cached reads, and a few are free. New users get a voucher. A minimum top-up is not documented. |
| OpenAI-compatible | Yes | Yes. Base URL https://api.novita.ai/openai, Chat Completions and Completions, streaming and non-streaming. |
| Anthropic-compatible | Yes, /v1/messages (verified with Claude Code) | Yes, for a subset of models, at https://api.novita.ai/anthropic, with a documented Claude Code setup guide. |
| Uncensored / abliterated | Five uncensored models, the Xploded line: GLM 5.3, GLM 5.3 Flash, DeepSeek V4.1 Flash, MiMo V2.6 Flash and Qwen3.8 27B. Every one calls tools. | A few older roleplay finetunes: its model list describes Llama 3 70B Euryale v2.1 as uncensored and good for roleplay, and also carries Euryale v2.2, L3 8B Lunaris and MythoMax L2 13B, with 4K to 8K context. Its terms prohibit using the service to get around safety filters. |
| Prompt retention | No prompt or completion stored; model, tokens and cost kept 90 days; signed prompt-hash receipt | Novita's terms and Trust Center describe zero data retention by default: content is not used to train its models or kept beyond generating the output, apart from automated safety screening and legal requirements. Technical metadata is kept for 6 months. |
| Per-key spend caps | Lifetime and monthly per-key caps, enforced by reserving the worst case and clamping max_tokens | Yes, on team accounts: a fixed or monthly spending cap per member and per API key, with the key paused when the cap is used up. Keys can also be limited to certain models and source IPs. Rate limits are per model, by tier. |
Where Novita AI is the better choice
- A much wider catalogue (DeepSeek, Qwen, Kimi, Llama and more, plus image, video, audio, batch and GPU instances) at low per-token prices, with 50% off batch jobs. MagmaRouter offers four models.
- Per-key budgets with model and IP allow-lists on team accounts, and documented no-training, no-retention defaults.
Where MagmaRouter is different
- Every MagmaRouter model is an uncensored Xploded model that can call tools: GLM 5.3 Xploded, GLM 5.3 Flash Xploded, DeepSeek V4.1 Flash Xploded, MiMo V2.6 Flash Xploded and Qwen3.8 27B Xploded, built to drive Claude Code, Cline and OpenCode. Novita's uncensored options are older roleplay finetunes with 4K to 8K context.
- MagmaRouter's per-key lifetime and monthly caps work on every account, not just teams, and are enforced before the call by reserving the worst-case cost and clamping max_tokens. Pay as you go from a $5 top-up with no subscription.
The uncensored models MagmaRouter sells
If you are comparing Novita AI because you want an uncensored or abliterated model API, these are the models on MagmaRouter: GLM 5.3 uncensored, GLM 5.3 Flash uncensored, DeepSeek V4.1 Flash uncensored, MiMo V2.6 Flash abliterated, Qwen3.8 27B uncensored and Qwen3.8 27B Fast uncensored. Every one calls tools, works in Claude Code through the Anthropic endpoint, and is priced per million tokens as listed on its model page.
| Model | What it is | Context | Price in / out | Supports |
|---|---|---|---|---|
| GLM 5.3 Xploded glm-5.3-xploded | GLM 5.3 uncensored API | 1M | $2.00 / $5.00 | tools, reasoning |
| GLM 5.3 Flash Xploded glm-5.3-flash-xploded | GLM 5.3 Flash uncensored API | 1M | $0.30 / $1.00 | tools, vision, reasoning |
| DeepSeek V4.1 Flash Xploded deepseek-v4.1-flash-xploded | DeepSeek V4.1 Flash uncensored API | 1M | $0.30 / $1.50 | tools, reasoning |
| MiMo V2.6 Flash Xploded mimo-v2.6-flash-xploded | MiMo V2.6 Flash abliterated API | 1M | $1.00 / $2.00 | tools, vision, reasoning |
| Qwen3.8 27B Xploded qwen3.8-27b-xploded | Qwen3.8 27B uncensored API | 1M | $1.50 / $5.00 | tools, vision, reasoning |
| Qwen3.8 27B Fast Xploded qwen3.8-27b-fast-xploded | Qwen3.8 27B Fast uncensored API | 524K | $1.50 / $5.00 | tools, vision, reasoning |
FAQ
Does Novita AI have uncensored models?
It lists a few older roleplay finetunes, such as Llama 3 70B Euryale v2.1, which Novita's own model page calls uncensored, plus Euryale v2.2, Lunaris and MythoMax, all with short 4K to 8K context windows. MagmaRouter's Xploded line is newer uncensored models that can call tools.
Can I use Novita AI with Claude Code?
Yes. Novita documents an Anthropic-compatible base URL and a Claude Code setup guide, though only some models support that endpoint. MagmaRouter also works with Claude Code through its /v1/messages endpoint.
Can I cap spending per API key on Novita AI?
Yes, on a team account: a fixed or monthly budget per key and per member, and the key pauses when the budget runs out. Personal accounts cannot use it.
Which uncensored models does MagmaRouter sell?
MagmaRouter sells GLM 5.3 Xploded, GLM 5.3 Flash Xploded, DeepSeek V4.1 Flash Xploded, MiMo V2.6 Flash Xploded, Qwen3.8 27B Xploded, Qwen3.8 27B Fast Xploded. The table above has the price and context for each.
Are these models abliterated or uncensored?
MiMo V2.6 Flash Xploded is abliterated: the refusal direction was removed from the weights. GLM 5.3 Xploded, GLM 5.3 Flash Xploded, DeepSeek V4.1 Flash Xploded, Qwen3.8 27B Xploded, Qwen3.8 27B Fast Xploded are uncensored: the refusal was trained back out. The abliterated vs uncensored page explains the difference.
Which one answers fastest?
MiMo V2.6 Flash Xploded usually starts answering in one to two seconds. The GLM 5.3 Flash model thinks before it answers, so its first word can take much longer.