MiniMax
MiniMax-M3
$0.4125 / 1M tokens
cheapest at W&B Inference (CoreWeave) · $0.23 in · $0.96 out
Wins 1 of 4
Head to head · prices updated Nov 1, 2026
MiniMax-M3 vs Kimi K3
Kimi K3 costs 13.1× as much as MiniMax-M3 at their cheapest providers (blended 3 input : 1 output tokens).
Moonshot AI Kimi
Kimi K3
$5.40 / 1M tokens
cheapest at SiliconFlow · $2.70 in · $13.50 out
Wins 2 of 4
- Lower priceMiniMax-M3
- Higher accuracy (GPQA)Kimi K3
- Longer contextTie
- More providersKimi K3
Verdict
MiniMax-M3 is the better deal on price
Kimi K3 costs 13.1× as much as MiniMax-M3 at their cheapest providers (blended 3 input : 1 output tokens).
Kimi K3 scores higher on GPQA Diamond: 93.1% (max reasoning) vs 90.9%.
So the choice is a trade-off: Kimi K3 for the hardest reasoning, MiniMax-M3 when cost matters more than the last few points of accuracy.
Your monthly bill
What would each cost you?
MiniMax-M3 saves you $49.78 a month ($597.36 a year) for this workload.
Side by side
| MiniMax-M3 | Kimi K3 | |
|---|---|---|
| Cheapest blended price / 1M | $0.4125 | $5.40 |
| Input / output per 1M | $0.23 / $0.96 | $2.70 / $13.50 |
| Accuracy (GPQA Diamond) | 90.9% | 93.1% (max reasoning) |
| Max context | 1M | 1M |
| Providers we track | 7 | 9 |
Standard list prices at each model's cheapest provider we track; blended = 3 input : 1 output tokens. GPQA Diamond from Epoch AI (CC BY 4.0), best measured reasoning setting, named next to each score. Methodology.
Where to buy each one
MiniMax-M3
- W&B Inference (CoreWeave)$0.4125
- DeepInfra$0.485
- Fireworks AI$0.525
- Novita AI$0.525
- SiliconFlow$0.525
Kimi K3
- SiliconFlow$5.40
- DeepInfra$5.70
- Alibaba Cloud Model Studio$6.00
- Moonshot AI$6.00
- Baseten$6.00
Questions people ask
Is MiniMax-M3 cheaper than Kimi K3?
Kimi K3 costs 13.1× as much as MiniMax-M3 at their cheapest providers (blended 3 input : 1 output tokens).
Which is more accurate, MiniMax-M3 or Kimi K3?
On GPQA Diamond (graduate-level science questions, measured by Epoch AI) MiniMax-M3 scores 90.9% and Kimi K3 93.1% (max reasoning). One benchmark is not the whole picture; test both on your own prompts.
How much does MiniMax-M3 cost per 1M tokens?
From $0.23 input and $0.96 output per 1M tokens at W&B Inference (CoreWeave), the cheapest of 7 providers we track.
How much does Kimi K3 cost per 1M tokens?
From $2.70 input and $13.50 output per 1M tokens at SiliconFlow, the cheapest of 9 providers we track.
Swap a model