MiniMax Token Plan Ultra logo

MiniMax Token Plan Ultra

LLM 3.3/10

~12.5B MiniMax-M3 tokens/mo; 6-7 concurrent agents; video gen 5 clips/day; same OpenAI + Anthropic endpoints

Tokens / month
1.7B

MiniMax Token Plan Ultra scores 3.3/10 for free value in LLM. The main constraint: Paid — $120/mo (~$1,200/yr annual). ~12.5B MiniMax-M3 tokens/mo; 6-7 concurrent agents; video gen 5 clips/day; same OpenAI + Anthropic endpoints.

Category LLM Kind Paid plan Caveats Quota shared across text/image/speech/music, gated by BOTH 5h-rolling AND weekly windows, no rollover; 'designed for individual, interactive developer use' — MiniMax recommends pay-as-you-go for production; peak-hour dynamic throttling (weekdays 15:00-17:30 UTC+8); big token numbers assume high cache-hit rates

Free value score

3.3 / 10
Show the math
Model Quality 8/10
value = 44 artificial_analysis_score
weighted score = 0.2 * 8/10
Access 10/10
value = 10/10
weighted score = 0.1 * 10/10
Price Efficiency 1/10
value = 0.01 value/$
weighted score = 0.7 * 1/10

Key metrics

MiniMax M3 44 Best model
1.7B T/mo Free tokens
0.0083x Value ratio

What you get

Paid — $120/mo (~$1,200/yr annual). ~12.5B MiniMax-M3 tokens/mo; 6-7 concurrent agents; video gen 5 clips/day; same OpenAI + Anthropic endpoints

~12.5B M3 tokens/mo for $120 => ~104M tokens per $1 — the largest flat open-model token pool of any subscription surveyed

Gotchas

No gotchas documented yet. Contribute one if you know a constraint we missed.

Sources

paid value llm-providers
Open service ← All LLM

Alternatives

Same category · similar quota
Canopy Wave 'Unlimited' Token PlanCanopy Wave 'Unlimited' Token Plan · paid · Paid plan
3.5

Kimi K2.6 + MiniMax M2.5: high-speed inference up to 50M tokens/mo, then low-priority throttled requests until next cycle; OpenAI-compatible API

0.06×42.8 AA0.1×
May be outdatedChecked 23d ago
Paid only
Cerebras Code ProCerebras Code Pro · paid · Paid plan
3.0

SOLD OUT (Jul 2026). Was: up to 24M tokens/day, 1M tokens/min at up to ~3,000 tok/s on Cerebras wafer-scale hardware; GLM 4.7-era open models

0.02×33.8 AA0.0×
May be outdatedChecked 23d ago
Paid only⛔ Sold out
StepFun Step Plan PlusStepFun Step Plan Plus · paid · Paid plan
3.6

step-3.7-flash (198B MoE / 11B active, native image+video input, 3 reasoning levels) at 400 prompts/5h (~6,000 model calls), 1,600 prompts/week; speed-first serving, no paid 'turbo' tier needed; OpenAI-compatible dedicated endpoint

0.1×29.7 AA0.1×
May be outdatedChecked 23d ago
Paid only