MiniMax H3 Benchmarks, Pricing & Rankings
MiniMax · 2026-08-28 · Open · Current
Snapshot 2026-10-14 · updated weekly
MiniMax H3 ranks #15 of 42 on the AMR leaderboard.
MiniMax H3 benchmark scores
| GPQA Diamond | 85.8% | |
| MMLU-Pro | 83.2% | |
| SWE-Bench Verified | 68.9% | |
| AIME 2025 | 87.9% | |
| LiveCodeBench v6 | 76.1% | |
| Arena Elo | 1412 |
MiniMax H3 price, speed & context
Input price
$0.60
per 1M tokens
Output price
$2.20
per 1M tokens
Speed
121
tok/s
Context
1M
tokens
About MiniMax H3
MiniMax H3 pairs an open-weights 1M context with multilingual strength, and its API pricing makes it the cheapest 1M-context model in the top 20.
Models around MiniMax H3 in the rankings
More MiniMax models
MiniMax H3 FAQ
Where does MiniMax H3 rank?
MiniMax H3 ranks #15 of 42 tracked AI models on the AMR aggregate as of October 2026, with an aggregate score of 88.2. Its strongest dimension is AIME 2025 at 87.9%.
How much does MiniMax H3 cost?
MiniMax H3 costs $0.60 per million input tokens and $2.20 per million output tokens on the standard API. At a typical 1:3 input/output agent workload, that blends to roughly $1.80 per million tokens.
What is the MiniMax H3 context window?
MiniMax H3 supports a 1M token context window and generates at a median 121 output tokens per second. That context fits roughly 667 standard documents in a single prompt.
Is MiniMax H3 open source?
Yes — MiniMax H3 ships open weights that you can download and self-host or serve through any compatible inference provider.
What is better than MiniMax H3?
DeepSeek Harness ranks one place higher at #14 with an AMR score of 88.8. MiniMax H3 still wins 0 of the tracked dimensions head-to-head — check the comparison table to see exactly which.