MiniMax M2 minimax-m2
minimax · minimax reasoning agentic efficient
#30 overall #13 agentic
- params
- 229B (A10B)
- arch
- moe
- context
- 200k
- license
- mit open weights
- released
- Oct 2025
- reasoning
- yes
- downloads/30d
- 232.3k
# architecture
- layers
- 62
- d_model
- 3072
- heads
- 48q / 8kv
- head dim
- 128
- experts
- top-8 of 256
- vocab
- 200064
- family
- minimax_m2
signal path of one layer, generated from the registry's structured fields — dimension lines quote the model's real numbers; an MoE trace forks at the router, a dense trace runs straight through. Geometry fields come from the repo's config.json.
# why ranked
Overall: #30 — score 56.7 ● high — all signals present
| signal | weight | input (0–100) |
|---|---|---|
| aa_intelligence | 0.70 | 60.3 |
| bench_composite | 0.30 | 48.2 |
benchmark panel evidence:
| complete panel | score (0–100) |
|---|---|
| aime-2025-v1 | 42.8 |
| browsecomp-v1 | 49.2 |
| swe-bench-verified-v1 | 52.5 |
Coding: provisional — score 44.4 ● low — a single signal; treat with caution
provisional: too few independent signals for a numbered position — the score is shown, but this model sorts after every ranked model.
| signal | weight | input (0–100) |
|---|---|---|
| aa_coding | 0.30 | — |
| aider_polyglot | 0.30 | — |
| swe_bench_verified | 0.40 | 44.4 |
missing signals are dropped and the remaining weights renormalized — never imputed.
Agentic: #13 — score 44.4 ● low — a single signal; treat with caution
| signal | weight | input (0–100) |
|---|---|---|
| swe_bench_verified | 0.60 | 44.4 |
| swe_rebench | 0.40 | — |
missing signals are dropped and the remaining weights renormalized — never imputed.
# trend
methodology changed during this history window; score movement across that boundary is not model movement. See methodology v5.
# benchmarks
| benchmark | score | source | date |
|---|---|---|---|
| AA Intelligence Index | 21.7 | Artificial Analysis | — |
| AA Math Index | 78.3 | Artificial Analysis | — |
| Aa-index | 0.6 / 1 | LLM Stats | — |
| AIME 2025 | 0.8 / 1 | LLM Stats | — |
| Browsecomp | 0.4 / 1 | LLM Stats | — |
| Browsecomp-zh | 0.5 / 1 | LLM Stats | — |
| SWE-bench Verified | 61.0 | SWE-bench | Nov 2025 |
| LMArena Elo (Code) | 1297.0 | LMArena | Aug 2026 |
# where to run
| provider | quant | ctx | $/M in | $/M out | $/M cache | price src | tps | uptime | ✓ |
|---|---|---|---|---|---|---|---|---|---|
| Featherless | — | — | — | — | via hfrouter | — | — | ||
| Amazon Bedrock | — | $0.15 | $0.60 | — | bedrock | — | — | ✓ | |
| Minimax | fp8 | 200k | $0.26 | $1.02 | — | via openrouter | — | 99.7% | |
| unknown | 192k | $0.30 | $1.20 | — | via openrouter | — | 100.0% | ||
| Novita | fp8 | 200k | $0.30 | $1.20 | — | novita | — | 99.9% | ✓ |
sorted by blended price ((3·input + output) / 4 per 1M) · ✓ = the provider's own catalog confirms the offer · "via …" prices are what the aggregator routing the offer charges, not the provider's own list price
# variants
source aliases
- aa
minimax-m2- arena
MiniMax-M2- openrouter
minimax/minimax-m2