Kimi K2-Thinking-0905
Kimi K2 Thinking is the latest, most capable version of open-source thinking model. Starting with Kimi K2, it is built as a thinking agent that reasons step-by-step while dynamically invoking tools. It sets a new state-of-the-art on Humanity's Last Exam (HLE), BrowseComp, and other benchmarks by dramatically scaling multi-step reasoning depth and maintaining stable tool-use across 200–300 sequential calls. At the same time, K2 Thinking is a native INT4 quantization model with 256k context window, achieving lossless reductions in inference latency and GPU memory usage. Key features include deep thinking & tool orchestration with end-to-end training to interleave chain-of-thought reasoning with function calls, native INT4 quantization via Quantization-Aware Training (QAT) achieving lossless 2x speed-up, and stable long-horizon agency maintaining coherent goal-directed behavior across up to 200–300 consecutive tool invocations.
Identité
| Organisation | Moonshot AI |
|---|---|
| Pays | — |
| Famille | — |
| Type | llm |
| Date de sortie | 2025-09-05 |
| Licence | mit |
| Licence commerciale | oui |
| Licence open source | oui |
| Poids ouverts | oui |
| Auto-hébergement (DGX Spark) | non — 1e+03 Md > 107 Md |
| Modalités | text |
| Fenêtre de contexte | 256 000 |
| KV cache / token (Q8) | — |
| Paramètres totaux | 1 T |
| Paramètres actifs (session) | 32 B |
| Source des paramètres actifs | web : https://huggingface.co/moonshotai/Kimi-K2-Thinking-0905 |
| Curation des paramètres actifs | 2026-11-04 |
| Prix par session | — |
| 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) | — |
| Tokens d'entraînement | 15 500 000 000 000 |
| Coupure de connaissance | — |
| Créé le | 2026-05-07T16:53:40.849758Z |
| Mis à jour le | 2026-11-04T06:05:14.871756Z |
| Source | LLM Stats |
| URL | https://llm-stats.com/models/kimi-k2-thinking-0905 |
Prix par fournisseur
Statut du modèle (offres + auto-hébergement) : non — hors DGX Spark (> 107 Md).
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 « — ».
Aucune donnée.
Souveraineté qualifiée par offre (modèle × fournisseur) ; survolez le badge pour les critères.
SouverainetéScores par catégorie (top_scores)
| Catégorie | Score |
|---|---|
| language | 0.90 |
| legal | 0.80 |
| math | 0.80 |
| physics | 0.60 |
| reasoning | 0.70 |
| search | 0.70 |
| finance | 0.80 |
| frontend_development | 0.70 |
| general | 0.70 |
| healthcare | 0.70 |
| agents | 0.50 |
| biology | 0.60 |
| chemistry | 0.60 |
| code | 0.50 |
| communication | 0.70 |
| creativity | 0.70 |
| vision | 0.50 |
| writing | 0.70 |
Scores par benchmark
| Benchmark | Catégorie | Score | Max | Normalisé (API) | Vérifié | Auto-déclaré | Daté du |
|---|---|---|---|---|---|---|---|
| aime-2025 | math | 1.00 | 1.00 | 1.00 | non | oui | 2026-11-04T06:05:14.888919Z |
| browsecomp | reasoning | 0.60 | 1.00 | 0.60 | non | oui | 2026-11-04T06:05:14.916310Z |
| browsecomp-zh | reasoning | 0.62 | 1.00 | 0.62 | non | oui | 2026-11-04T06:05:14.919414Z |
| finsearchcomp-t3 | finance | 0.47 | 1.00 | 0.47 | non | oui | 2026-11-04T06:05:14.951398Z |
| frames | reasoning | 0.87 | 1.00 | 0.87 | non | oui | 2026-11-04T06:05:14.922210Z |
| gpqa | reasoning | 0.84 | 1.00 | 0.84 | non | oui | 2026-11-04T06:05:14.901945Z |
| healthbench | healthcare | 0.58 | 1.00 | 0.58 | non | oui | 2026-11-04T06:05:14.913466Z |
| hmmt-2025 | math | 0.97 | 1.00 | 0.97 | non | oui | 2026-11-04T06:05:14.897709Z |
| humanity's-last-exam | reasoning | 0.51 | 1.00 | 0.51 | non | oui | 2026-11-04T06:05:14.880566Z |
| imo-answerbench | math | 0.79 | 1.00 | 0.79 | non | oui | 2026-11-04T06:05:14.945757Z |
| livecodebench-v6 | reasoning | 0.83 | 1.00 | 0.83 | non | oui | 2026-11-04T06:05:14.937016Z |
| mmlu-pro | language | 0.85 | 1.00 | 0.85 | non | oui | 2026-11-04T06:05:14.904738Z |
| mmlu-redux | language | 0.94 | 1.00 | 0.94 | non | oui | 2026-11-04T06:05:14.907579Z |
| multi-swe-bench | reasoning | 0.42 | 1.00 | 0.42 | non | oui | 2026-11-04T06:05:14.931339Z |
| ojbench | reasoning | 0.49 | 1.00 | 0.49 | non | oui | 2026-11-04T06:05:14.939933Z |
| scicode | reasoning | 0.45 | 1.00 | 0.45 | non | oui | 2026-11-04T06:05:14.934217Z |
| seal-0 | reasoning | 0.56 | 1.00 | 0.56 | non | oui | 2026-11-04T06:05:14.948645Z |
| swe-bench-multilingual | reasoning | 0.61 | 1.00 | 0.61 | non | oui | 2026-11-04T06:05:14.928505Z |
| swe-bench-verified | reasoning | 0.71 | 1.00 | 0.71 | non | oui | 2026-11-04T06:05:14.925425Z |
| terminal-bench | reasoning | 0.47 | 1.00 | 0.47 | non | oui | 2026-11-04T06:05:14.942797Z |
| writingbench | writing | 0.74 | 1.00 | 0.74 | non | oui | 2026-11-04T06:05:14.910425Z |
Classements
| Catégorie | Rang | Score | Note conservatrice | Benchmarks évalués | Méthode | Daté du | Provenance |
|---|---|---|---|---|---|---|---|
| math | 33 | 0.80 | 37.75 | 7 | trueskill | 2026-11-04T08:08:42.442591Z | rankings |
| agents | 93 | — | 16.10 | 2 | trueskill | 2026-10-19 | model_history |
| code | 107 | — | 20.42 | 5 | 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