MiniMax M3
MiniMax M3 is the first open-weight model to combine three frontier capabilities: top-tier coding and agentic performance, a 1M-token context window, and native multimodality. It is powered by MiniMax Sparse Attention (MSA), a new sparse attention architecture that partitions the KV cache into blocks to cut per-token compute at long context — roughly 1/20 the cost of the previous generation at 1M tokens, with more than 9x faster prefill and more than 15x faster decode while matching full attention on most capabilities. Trained with mixed-modality data from step zero across 100T+ tokens, M3 natively supports image and video input and can operate a desktop computer. On SWE-Bench Pro it scores 59.0%, surpassing GPT-5.5 and Gemini 3.1 Pro and approaching Opus 4.7, and on BrowseComp it scores 83.5%, surpassing Opus 4.7. M3 supports toggling thinking on or off at request time.
Identité
| Organisation | MiniMax |
|---|---|
| Pays | — |
| Famille | — |
| Type | llm |
| Date de sortie | 2026-06-01 |
| Licence | mit |
| Licence commerciale | oui |
| Licence open source | oui |
| Poids ouverts | oui |
| Auto-hébergement (DGX Spark) | non — 428 Md > 107 Md |
| Modalités | image|text|video |
| Fenêtre de contexte | 1 048 576 |
| KV cache / token (Q8) | 61.4 ko/token |
| Paramètres totaux | 428 B |
| Paramètres actifs (session) | 23 B |
| Source des paramètres actifs | web : https://huggingface.co/MiniMaxAI/MiniMax-M3 |
| Curation des paramètres actifs | 2026-11-04 |
| Prix par session | 0.50 $ (1 000 000 tokens entrée + 200 000 sortie, prix le moins cher par fournisseur) |
| Tokens / tâche (Data Players) | 1.2 M |
| Tokens / tâche (Terminal-Bench) | 83.3 M (est.) |
| Ratio Terminal-Bench / Data Players | 69.4× |
| Coût / tâche (Terminal-Bench) | 4.20 $ (est.) |
| Tokens d'entraînement | — |
| Coupure de connaissance | — |
| Créé le | 2026-06-05T21:54:39.691299Z |
| Mis à jour le | 2026-11-04T06:04:55.316864Z |
| Source | LLM Stats |
| URL | https://llm-stats.com/models/minimax-m3 |
IA Act — preuves fournisseur
| Critère | Note / 5 | Poids |
|---|---|---|
| Documentation technique | 2,5 | 5 % |
| Résumé des données d'entraînement | 0,0 | 35 % |
| Conformité droit d'auteur | 0,0 | 35 % |
| Code of Practice GPAI | 1,2 | 20 % |
| Risque systémique | 0,0 | 5 % |
Curation : 2026-10-21.
IA ActPrix par fournisseur
Statut du modèle (offres + auto-hébergement) : souverain — via Arkane (cortecs), Tensorix (cortecs, eurouter).
Toutes sources confondues : catalogue llm-stats, routeurs EUrouter/Cortecs et catalogues directs (OVHcloud, Scaleway, Infomaniak). Les prix sont convertis en USD pour la comparaison ; le prix publié et le taux BCE utilisé sont dans l'infobulle. Les prix inconnus restent indiqués « — ».
| Fournisseur | Source | Souveraineté | Entrée (/M tok) | Sortie (/M tok) | Quantification | Contexte | Latence (s) | Débit (tok/s) | Statut |
|---|---|---|---|---|---|---|---|---|---|
| Arkane | cortecs | souverain | 0.30 $ | 1.21 $ | — | 1 000 000 | — | — | — |
| Tensorix | cortecs, eurouter | souverain | 0.40 $ | 1.99 $ | fp8 | 1 048 576 | — | — | — |
| DeepInfra | llm-stats | non | 0.28 $ | 1.10 $ | — | — | — | — | active |
| Fireworks AI | llm-stats | non | 0.30 $ | 1.20 $ | — | — | — | — | active |
| MiniMax | llm-stats | non | 0.30 $ | 1.20 $ | — | — | — | — | active |
| Novita | llm-stats | non | 0.30 $ | 1.20 $ | — | — | — | — | active |
| Together AI | llm-stats | non | 0.30 $ | 1.20 $ | — | — | — | — | active |
Souveraineté qualifiée par offre (modèle × fournisseur) ; survolez le badge pour les critères.
SouverainetéScores par catégorie (top_scores)
| Catégorie | Score |
|---|---|
| math | 18.40 |
| multimodal | 0.80 |
| productivity | 0.90 |
| reasoning | 2.60 |
| search | 0.80 |
| structured_output | 0.90 |
| finance | 0.60 |
| frontend_development | 0.80 |
| general | 2.00 |
| healthcare | 0.80 |
| agents | 0.60 |
| code | 0.50 |
| systems | 0.40 |
| tool_calling | 0.80 |
| vision | 0.80 |
Scores par benchmark
| Benchmark | Catégorie | Score | Max | Normalisé (API) | Vérifié | Auto-déclaré | Daté du |
|---|---|---|---|---|---|---|---|
| apex-agents | reasoning | 0.28 | 1.00 | 0.28 | non | oui | 2026-11-04T06:04:55.475895Z |
| bankertoolbench | agents | 0.76 | 1.00 | 0.76 | non | oui | 2026-11-04T06:04:55.432994Z |
| browsecomp | reasoning | 0.84 | 1.00 | 0.84 | non | oui | 2026-11-04T06:04:55.422990Z |
| cl-bench | code | 0.20 | 1.00 | 0.20 | non | oui | 2026-11-04T06:04:55.378102Z |
| claw-eval | code | 0.74 | 1.00 | 0.74 | non | oui | 2026-11-04T06:04:55.479441Z |
| draco | agents | 0.73 | 1.00 | 0.73 | non | oui | 2026-11-04T06:04:55.428428Z |
| finance-agent-v2 | agents | 0.48 | 1.00 | 0.48 | non | non | 2026-11-04T06:04:55.533033Z |
| frontiercode-1.1 | reasoning | 0.15 | 1.00 | 0.15 | non | non | 2026-11-04T06:04:55.543260Z |
| gdpval-rubrics | general | 0.75 | 1.00 | 0.75 | non | oui | 2026-08-11T05:53:37.001135Z |
| imo-2025 | math | 35.00 | 42.00 | 0.83 | non | oui | 2026-11-04T06:04:55.523250Z |
| kernelbench-hard | code | 0.29 | 1.00 | 0.29 | non | oui | 2026-11-04T06:04:55.413138Z |
| livebench | math | 0.70 | 1.00 | 0.70 | non | non | 2026-11-04T06:04:55.536301Z |
| livesqlbench | code | 0.40 | 1.00 | 0.40 | non | oui | 2026-11-04T06:04:55.370380Z |
| loca-bench-256k | agents | 0.49 | 1.00 | 0.49 | non | oui | 2026-11-04T06:04:55.468943Z |
| mcp-atlas | reasoning | 0.74 | 1.00 | 0.74 | non | oui | 2026-11-04T06:04:55.472585Z |
| mmmu-pro | vision | 0.78 | 1.00 | 0.78 | non | oui | 2026-11-04T06:04:55.503852Z |
| nl2repo | code | 0.42 | 1.00 | 0.42 | non | oui | 2026-11-04T06:04:55.353087Z |
| officeqa-pro | agents | 0.45 | 1.00 | 0.45 | non | oui | 2026-11-04T06:04:55.443082Z |
| omnidocbench-1.5 | vision | 0.92 | 1.00 | 0.92 | non | oui | 2026-11-04T06:04:55.493131Z |
| osworld-verified | multimodal | 0.70 | 1.00 | 0.70 | non | oui | 2026-11-04T06:04:55.482959Z |
| paperbench | code | 0.53 | 1.00 | 0.53 | non | oui | 2026-11-04T06:04:55.418364Z |
| posttrainbench | agents | 0.37 | 1.00 | 0.37 | non | oui | 2026-11-04T06:04:55.403086Z |
| spreadsheetbench-v1 | agents | 0.89 | 1.00 | 0.89 | non | oui | 2026-11-04T06:04:55.453049Z |
| svg-bench | code | 0.64 | 1.00 | 0.64 | non | oui | 2026-11-04T06:04:55.393066Z |
| swe-atlas-codebase-qna | code | 0.38 | 1.00 | 0.38 | non | oui | 2026-11-04T06:04:55.340074Z |
| swe-atlas-test-writing | code | 0.31 | 1.00 | 0.31 | non | oui | 2026-11-04T06:04:55.363249Z |
| swe-bench-pro | reasoning | 0.59 | 1.00 | 0.59 | non | oui | 2026-11-04T06:04:55.329262Z |
| swe-bench-verified | reasoning | 0.81 | 1.00 | 0.81 | non | oui | 2026-11-04T06:04:55.323414Z |
| swe-fficiency | code | 0.35 | 1.00 | 0.35 | non | oui | 2026-11-04T06:04:55.367181Z |
| terminal-bench-2.1 | reasoning | 0.66 | 1.00 | 0.66 | non | oui | 2026-11-04T06:04:55.333055Z |
| usamo-2026 | math | 36.00 | 42.00 | 0.86 | non | oui | 2026-11-04T06:04:55.529115Z |
| vibe-v2 | code | 0.50 | 1.00 | 0.50 | non | oui | 2026-11-04T06:04:55.381814Z |
| video-mme | multimodal | 0.85 | 1.00 | 0.85 | non | oui | 2026-11-04T06:04:55.518208Z |
| videommmu | multimodal | 0.85 | 1.00 | 0.85 | non | oui | 2026-11-04T06:04:55.513318Z |
| yc-bench | agents | 2 100 000.00 | 10 000 000.00 | — | non | oui | 2026-11-04T06:04:55.462498Z |
Classements
| Catégorie | Rang | Score | Note conservatrice | Benchmarks évalués | Méthode | Daté du | Provenance |
|---|---|---|---|---|---|---|---|
| multimodal | 44 | 0.80 | 28.59 | 5 | trueskill | 2026-11-04T08:08:46.488887Z | rankings |
| agents | 61 | — | 25.42 | 17 | trueskill | 2026-10-19 | model_history |
| code | 53 | — | 30.88 | 11 | trueskill | 2026-10-19 | model_history |
Données
Dates de curation (récupération depuis l'API) : models 2026-11-04 · benchmarks 2026-11-04 · scores 2026-11-04 · updates 2026-11-04 · account 2026-11-04 · model_history 2026-11-04 · rankings 2026-11-04