MagmaRouter vs OrcaRouter

MagmaRouter and OrcaRouter

OrcaRouter is a hosted router over about 200 mostly frontier models, with adaptive routing and a $0-markup pitch. On price and routing features it leads; MagmaRouter’s difference is the uncensored catalogue and the hard cap.

Facts checked September 2026 against OrcaRouter’s own site, docs and policies.

MagmaRouterOrcaRouter
What it isHosted gateway for one uncensored, tool capable model line (Xploded).A hosted OpenAI- and Anthropic-compatible router over ~200 models, with adaptive routing, failover, guardrails and an MIT-licensed self-hostable "Lite" edition.
PricingPay as you go, $5 minimum top-up, no subscription required.Provider list price with $0 routing fee on the free Hacker tier; a Team plan (price inconsistent on their page); BYOK carries a 5% default fee.
OpenAI-compatibleYesYes
Anthropic-compatibleYes, /v1/messages (verified with Claude Code)Yes, /v1/messages
Uncensored / abliteratedSeven uncensored models, the Xploded line: GLM 5.3, GLM 5.3 Flash, DeepSeek V4.1 Flash 1017, DeepSeek V4.1 Flash Lite 1017, MiMo V2.6 Flash, Qwen3.8 27B and Qwen3.8 27B Ultrafast. Every one calls tools.None: ~200 mostly frontier/commercial models, nothing positioned as uncensored
Prompt retentionNo prompt or completion stored; model, tokens and cost kept 90 days; signed prompt-hash receiptNo prompt/output content retained (default, not opt-in); usage metadata kept 13 months including source IP; not used for training
Per-key spend capsLifetime and monthly per-key caps, enforced by reserving the worst case and clamping max_tokensPer-key total USD credit limit and a key expiry; no per-key rate limit or model allowlist; monthly/in-flight mechanics not verified

Where OrcaRouter is the better choice

  • Price: provider list price with $0 markup on the free tier, plus BYOK at a 5% default fee.
  • Features: adaptive routing that grades each prompt and picks a model, failover, guardrails and an agent firewall, a prompt registry, and a self-hostable MIT "Lite" edition.

Where MagmaRouter is different

  • Catalogue shape: MagmaRouter sells a short line of uncensored models that all call tools; OrcaRouter is ~200 mostly frontier models with none positioned as uncensored.
  • Retention and caps: MagmaRouter keeps metadata 90 days (OrcaRouter keeps usage metadata 13 months, including source IP) and has both lifetime and monthly per-key caps with worst-case reservation, versus a single total USD credit limit.

The uncensored models MagmaRouter sells

If you are comparing OrcaRouter 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 1017 abliterated, DeepSeek Flash Lite abliterated, MiMo V2.6 Flash abliterated, Qwen3.8 27B uncensored and Qwen3.8 27B Ultrafast 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.

ModelWhat it isContextPrice in / outSupports
GLM 5.3 Xploded
glm-5.3-xploded
GLM 5.3 uncensored API1M$2.00 / $5.00tools, reasoning
GLM 5.3 Flash Xploded
glm-5.3-flash-xploded
GLM 5.3 Flash uncensored API1M$0.30 / $1.00tools, vision, reasoning
DeepSeek V4.1 Flash 1017 Xploded
deepseek-v4.1-flash-1017-xploded
DeepSeek V4.1 Flash 1017 abliterated API1M$1.25 / $2.50tools, vision, reasoning
DeepSeek Flash Lite Xploded
deepseek-v4.1-flash-lite-1017-xploded
DeepSeek Flash Lite abliterated API512K$0.36 / $1.44tools, vision, reasoning
MiMo V2.6 Flash Xploded
mimo-v2.6-flash-xploded
MiMo V2.6 Flash abliterated API1M$1.00 / $2.00tools, vision, reasoning
Qwen3.8 27B Xploded
qwen3.8-27b-xploded
Qwen3.8 27B uncensored API1M$1.50 / $5.00tools, vision, reasoning
Qwen3.8 27B Ultrafast Xploded
qwen3.8-27b-fast-xploded
Qwen3.8 27B Ultrafast uncensored API524K$1.75 / $5.50tools, vision, reasoning

FAQ

Which uncensored models does MagmaRouter sell?

MagmaRouter sells GLM 5.3 Xploded, GLM 5.3 Flash Xploded, DeepSeek V4.1 Flash 1017 Xploded, DeepSeek Flash Lite Xploded, MiMo V2.6 Flash Xploded, Qwen3.8 27B Xploded, Qwen3.8 27B Ultrafast Xploded. The table above has the price and context for each.

Are these models abliterated or uncensored?

DeepSeek V4.1 Flash 1017 Xploded and DeepSeek Flash Lite Xploded and MiMo V2.6 Flash Xploded are abliterated: the refusal direction was removed from the weights. GLM 5.3 Xploded, GLM 5.3 Flash Xploded, Qwen3.8 27B Xploded, Qwen3.8 27B Ultrafast 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.

Other comparisons