Massive context faceoff: MiniMax-01 (4M tokens) vs Google Gemini 3.1 Pro (2M tokens).
MiniMax-01 · minimax
gemini-3.1-pro · google
Benchmark
| Workload Scenario | MiniMax-01 (4M Context) | Gemini 3.1 Pro | Price Delta |
|---|---|---|---|
| 1M input tokens (raw text) | $0.20 | $2.00 | −$1.80 |
| 1M output tokens (generation) | $1.10 | $12.00 | −$10.90 |
| 1M tokens · 70% input / 30% output mix | $0.47 | $5.00 | −$4.53 |
| Standard chat turn (4K in / 800 out, 50% cached) | $0.00136 | $0.014 | −$0.0126 |
| Monthly scale (10K requests / day) | $408.00 | $4,200.00 | −$3,792.00 |
Negative difference = MiniMax-01 (4M Context) is cheaper. Positive = Gemini 3.1 Pro is cheaper.
Scaling Curve
Total cost of a token volume at 70/30 input/output split (uncached). Log-log scale.
Verdict
MiniMax-01 (4M Context) is cheaper on both input and output rates, so it costs less at every input/output mix. Price alone still isn't the whole decision: capability, latency and context limits (MiniMax-01 (4M Context): 4M, Gemini 3.1 Pro: 2M) may justify the premium for your task.
FAQ
On input, MiniMax-01 (4M Context) is cheaper ($0.20/M vs $2.00/M). On output, MiniMax-01 (4M Context) is cheaper ($1.10/M vs $12.00/M). The same model is cheaper on both sides, so it wins at every mix.
A chat-style request (4,000 input + 800 output tokens, 50% cached) costs $0.00136 on MiniMax-01 (4M Context) and $0.014 on Gemini 3.1 Pro — MiniMax-01 (4M Context) is 10.3× cheaper for that workload.
MiniMax-01 (4M Context) supports 4,000,000 tokens (64K max output); Gemini 3.1 Pro supports 2,000,000 (128K max output). MiniMax-01 (4M Context) fits 2.0× more context, which matters for long documents and agents.