Données croisées (Artificial Analysis Intelligence Index v4.1.1, BenchLM, LiveBench, OpenRouter). Cliquez sur les en-têtes pour trier.
| Modèle ↕ | Compagnie ↕ | Date ↕ | Params ↕ | Params Actifs ↕ | Prix In $/M ↕ | Prix Out $/M ↕ | Compétences ↕ | Indice Intel /63 ↕ | Indice Code /100 ↕ | Coût/task AA ($) ↕ |
|---|---|---|---|---|---|---|---|---|---|---|
| Claude Opus 5 (Max) | Anthropic | 2026-07 | Propriétaire | - | $15.00 | $75.00 | Reasoning, Coding, Agentic | 63 | 84.5 | $0.42 |
| Claude Fable 5 | Anthropic | 2026-08 | Propriétaire | - | $8.00 | $24.00 | Multimodal, Agentic, LiveBench Lead | 62 | 83.0 | $0.31 |
| GPT-5.6 Sol (max) | OpenAI | 2026-07 | Propriétaire | - | $10.00 | $30.00 | Reasoning, Math, Code | 61 | 81.1 | $0.38 |
| Gemini 3.7 Flash High | 2026-08 | Propriétaire | - | $0.375 | $1.875 | Agentic, Speed, 1M Context | 59 | 78.8 | $0.08 | |
| Kimi K3 (max) | Moonshot AI | 2026-08 | 2.8T | 104B | $2.30 | $6.90 | Open Weights, Long Context, Reasoning | 60 | 79.2 | $0.19 |
| GLM-5.3 (max) | Z.AI | 2026-08 | 753B | 40B | $0.90 | $2.70 | Open Weights, Agentic, 1M Context | 60 | 78.0 | $0.15 |
| Qwen3.8 2.4T A95B | Alibaba | 2026-07 | 2.4T | 95B | $1.20 | $3.60 | Open Weights, Code, Math | 58 | 78.5 | $0.14 |
| Grok 4.6 | xAI | 2026-07 | Propriétaire | - | $5.00 | $15.00 | Real-time Web, Reasoning | 58 | 78.0 | $0.22 |
| DeepSeek V4 Pro 0813 | DeepSeek | 2026-08 | 1.6T | 49B | $0.70 | $2.10 | Open Weights, Code, Reasoning | 53 | 77.4 | $0.11 |
| Qwen3.8 27B (xhigh) | Alibaba | 2026-07 | 27B | 27B | $0.40 | $1.20 | Open Weights, Local Deploy, Coding | 52 | 72.5 | $0.05 |
| MiniMax-M3 | MiniMax | 2026-07 | 428B | 23B | $0.20 | $0.60 | Open Weights, High Speed (121 t/s) | 45 | 67.3 | $0.04 |
| GLM-5.3-Flash | Z.AI | 2026-08 | 320B | 18B | $0.10 | $0.30 | Open Weights, Ultra Efficient | 57 | 71.6 | $0.03 |