MiniMax · Benchmarks

MiniMax M3.1 Flash benchmarks

MiniMax M3.1 Flash has 2 published scores in this dataset, from a MiniMax release dated 2026-10-19. It does not yet have scores in enough categories to receive a composite index. Its strongest area is reasoning (index 67.4) and its weakest is coding (index 54.0). Its best placing is on Humanity's Last Exam, where its 43.4 ranks 12th of 23 models with a published score.

Open weightsOfficial page

Category profile

Index 0–100: each score relative to the best published score on that test.

ReasoningCodingAgenticKnowledge
Reasoning
67.41 test · not ranked
Coding
54.01 test · not ranked
Agentic
—no data
Knowledge
—no data

Model facts

Lab
MiniMax
Type
Open weights
Released
2026-10-19
Parameters
428B total, 23B active (MoE)
Context
1M
License
MiniMax Community License

MiniMax M3.1 Flash hardware requirements and cost

Even at 4-bit it needs about 295 GB (4× H100 (320 GB)), well beyond a home workstation, so most people run it on rented multi-GPU servers. Full BF16 precision needs about 984 GB (8× H200 (1,128 GB)).

Estimated GPU memory for MiniMax M3.1 Flash
PrecisionWeightsWith headroomSmallest setup
BF16 (full precision)856 GB984 GB8× H200 (1,128 GB)
FP8 / 8-bit428 GB492 GB4× H200 (564 GB)
4-bit (GGUF Q4_K_M)257 GB295 GB4× H100 (320 GB)

Estimate: parameters × bytes per parameter, plus 15% for the KV cache and runtime at modest context lengths. Long contexts and large batches need more.

Lighter or stronger open models

Open models ranked just above and below on the composite index, with their 4-bit memory needs.

MiniMax M3.1 Flash speed and pricing

Measured figures link to their source; estimates are labelled.

API output speed
136 tok/sArtificial Analysis median · Artificial Analysis
Time to first token
1.08 sOpenRouter model page provider stats
API price per 1M tokens
$0.150 in · $0.600 outlab's own endpoint on OpenRouter · OpenRouter endpoints API
Compare speed and price across all models

Score by test

Rank is among every model with a published score on that test.

Humanity's Last Exam43.4%no tools, text-only subset (2,158 q)#12 of 23 · 21.0 behind Claude Opus 5.5Terminal-Bench 4.038.1%Terminal-Bench 4.0, Codex 0.149.1 harness#11 of 16 · 32.5 behind Claude Sonnet 5.5

Nearby in the composite ranking

Models with the closest composite index.

  1. 1Claude Opus 5.598.4
  2. 2Kimi K3.1 Pro94.4
  3. 3Claude Sonnet 5.591.9
  4. 4GPT-6 Astra91.4
  5. 5Claude Fable 5.191.2
  6. 6DeepSeek V4 Pro91.0
  7. 7Kimi K390.6

Head-to-head comparisons

Side-by-side pages built from the tests both models report.

No head-to-head page for MiniMax M3.1 Flash yet. Compare it on the rankings dashboard.

Sources

ModelTestScoreSettingSource
MiniMax M3.1 FlashHumanity's Last Exam43.4no tools, text-only subset (2,158 q)Hugging Face model card (MiniMax-M3.1-Flash)
MiniMax M3.1 FlashTerminal-Bench 4.038.1Terminal-Bench 4.0, Codex 0.149.1 harnessHugging Face model card (MiniMax-M3.1-Flash)

Scores are published results from model cards, lab reports, and public leaderboards, each linked to its source. Settings such as thinking mode or test version differ between sources and are shown next to each score. OpenLLM Buddy catalog models are uncensored fine-tunes; their scores are those of the official base model. Data collected 2026-10-30. How the index works