MiniMax M3 vs GLM 5.2: API pricing comparison
MiniMax MiniMax M3 vs Zhipu GLM 5.2: token prices, batch and cache rates, Arena ELO, and cost per request for 8 business use cases. Prices in USD per 1 million tokens. Data snapshot: 2026-11-12.
MiniMax M3 vs GLM 5.2: the short version
At the prices we quote, neither is cheaper on both sides: GLM 5.2 costs less on input ($0.007 against $0.30 per 1M tokens, 98% less), while MiniMax M3 costs less on output ($1.20 against $6.80 per 1M tokens, 82% less). The price we quote for GLM 5.2 is an outlier listing: one host's rate with input near free and output at or above the going rate, not what most hosts charge, so a cost built on it can be far from what you would pay. With the Support Ticket profile (1,500 input and 500 output tokens per request), 1,000 requests cost $1.05 on MiniMax M3 and $3.41 on GLM 5.2 at list price, 69% less on MiniMax M3. On Arena ELO, GLM 5.2 scores 1469 and MiniMax M3 scores 1428, 41 points apart (scores as of 2026-11-08). The context window is 1M tokens on both. MiniMax M3 is open weights (the MiniMax Community License licence, which restricts some uses); GLM 5.2 is open source (MIT), so it can be self-hosted. Prices as quoted on 2026-11-12. Each price is the list price of one host, the one we quote for that model. A model sold by several hosts has other prices, listed on its model page.
MiniMax M3 pricing (MiniMax)
- Input: $0.30 per 1M tokens
- Output: $1.20 per 1M tokens
- Batch API (input/output): $0.224 / $0.88
- Prompt-cache read: $0.06
- Context window: 1M
- Arena ELO: 1428 (as of 2026-11-08)
- Full MiniMax M3 pricing page
GLM 5.2 pricing (Zhipu)
- Input: $0.007 per 1M tokens
- Output: $6.80 per 1M tokens
- Batch API (input/output): not published
- Prompt-cache read: $0.006
- Context window: 1M
- Arena ELO: 1469 (as of 2026-11-08)
- Full GLM 5.2 pricing page
Cost per 1,000 requests by business use case
List price vs after discounts (prompt caching on the cached input share, batch API for async workloads, only where the provider publishes those rates).
| Use case | Tokens (in/out) | MiniMax M3 list price | MiniMax M3 after discounts | GLM 5.2 list price | GLM 5.2 after discounts |
|---|---|---|---|---|---|
| Support Ticket | 1,500 / 500 | $1.05 | $0.83 | $3.41 | $3.41 |
| Knowledge Q&A | 2,000 / 800 | $1.56 | $1.22 | $5.45 | $5.45 |
| Meeting Summary | 10,000 / 1,200 | $4.44 | $3.13 | $8.23 | $8.23 |
| Marketing Content | 2,500 / 1,800 | $2.91 | $2.79 | $12.26 | $12.26 |
| Coding Task | 3,000 / 2,000 | $3.30 | $2.94 | $13.62 | $13.62 |
| Invoice Processing | 1,500 / 600 | $1.17 | $0.79 | $4.09 | $4.09 |
| Call Summary | 2,000 / 700 | $1.44 | $1.03 | $4.77 | $4.77 |
| Agent Workflow | 6,000 / 3,000 | $5.40 | $4.39 | $20.44 | $20.44 |
Common questions about MiniMax M3 vs GLM 5.2
- Which is cheaper, MiniMax M3 or GLM 5.2?
- At the prices we quote, neither is cheaper on both sides: GLM 5.2 costs less on input ($0.007 against $0.30 per 1M tokens, 98% less), while MiniMax M3 costs less on output ($1.20 against $6.80 per 1M tokens, 82% less). The price we quote for GLM 5.2 is an outlier listing: one host's rate with input near free and output at or above the going rate, not what most hosts charge, so a cost built on it can be far from what you would pay. Prices as quoted on 2026-11-12. Each price is the list price of one host, the one we quote for that model. A model sold by several hosts has other prices, listed on its model page.
- What do 1,000 Support Ticket requests cost on MiniMax M3 and on GLM 5.2?
- With the Support Ticket profile (1,500 input and 500 output tokens per request), 1,000 requests cost $1.05 on MiniMax M3 and $3.41 on GLM 5.2 at list price, 69% less on MiniMax M3. Prices as quoted on 2026-11-12.
- How do MiniMax M3 and GLM 5.2 compare on Arena score, context window and licence?
- On Arena ELO, GLM 5.2 scores 1469 and MiniMax M3 scores 1428, 41 points apart (scores as of 2026-11-08). The context window is 1M tokens on both. MiniMax M3 is open weights (the MiniMax Community License licence, which restricts some uses); GLM 5.2 is open source (MIT), so it can be self-hosted.
Open this comparison in the interactive compare tool: add up to two more models, copy a shareable link, or export the table as CSV.
These are list prices: real cost also depends on caching, batching, negotiated discounts, marketplace agreements, and existing contracts. Get a FinOps review from OptimNow or compare all models.