AI Models Database (2026)
Every major AI model — parameters, context window, pricing, licensing, benchmarks, and the hardware needed to run the open ones locally. Sortable and filterable.
API pricing by provider: Claude · Gemini · DeepSeek · Grok · AI benchmarks explained
Free open dataset: all 43 models (specs, pricing, VRAM) as CSV ou JSON structuré, regenerated 20 September 2026. Mirrors on GitHub et Hugging Face. Licensed CC BY 4.0 — free to use with attribution.
Cite this dataset:
Convly AI Models Database, Convly (2026). CC BY 4.0. https://convly.ai/models/| Modèle | Développeur | Type | Contexte | Entrée : $/1 million | Sortie : $/1 million | Ouvrir |
|---|---|---|---|---|---|---|
| Claude Fable 5 | Anthropic | LLM (raisonnement de pointe) | 1 million | $10.00 | $50.00 | |
| Claude Haiku 4.5 | Anthropic | LLM | 200 000 | $1.00 | $5.00 | |
| Claude Opus 4.8 | Anthropic | LLM (raisonnement) | 1 million | $5.00 | $25.00 | |
| Claude Opus 5 | Anthropic | Raisonnement / agentic | 1 million | $5.00 | $25.00 | |
| Claude Sonnet 4.6 | Anthropic | LLM | 1 million | $3.00 | $15.00 | |
| Claude Sonnet 5 | Anthropic | Général / agentic | 1 million | $2.00 | $10.00 | |
| DeepSeek R1 | DeepSeek | LLM (MoE, raisonnement) | 128 K | $0.50 | $2.15 | |
| DeepSeek R1 Distill Llama 70B | DeepSeek | LLM (dense, raisonnement) | 128 K | $0.80 | $0.80 | |
| DeepSeek V4-Flash | DeepSeek | LLM (architecture MoE) | 1 million | $0.14 | $0.28 | |
| DeepSeek V4-Pro | DeepSeek | LLM (architecture MoE) | 1 million | $0.435 | $0.87 | |
| Gemini 2.5 Pro | Google (Google DeepMind) | Modèle de langage volumineux (LLM) de pointe pour le raisonnement / la réflexion | 1 million (1 048 576 jetons) | $1.25 | $10 | |
| Gemini 3.1 Pro | LLM (multimodale) | 1,05 million | $2.00 | $12.00 | ||
| Gemini 3.5 Flash | LLM (multimodale) | 1 million | $1.50 | $9.00 | ||
| Gemini 3.6 Flash | Rapide / généraliste | 1 million | $1.50 | $7.50 | ||
| Gemma 3 12B | LLM (multimodale) | 128 K | $0.05 | $0.15 | ||
| Gemma 3 27B | LLM (multimodale) | 128 K | $0.08 | $0.16 | ||
| Gemma 3 4B | LLM (multimodale) | 128 K | $0.05 | $0.1 | ||
| GLM 5.2 | Zhipu AI | LLM (codage/agent, MoE) | 1 million | $1.4 | $4.4 | |
| GPT-5.5 | OpenAI | LLM (raisonnement) | 1,05 million | $5.00 | $30.00 | |
| GPT-5.6 Sol | OpenAI | Raisonnement / agentic | 1,05 million | $5.00 | $30.00 | |
| GPT-6 Astra | OpenAI | Raisonnement / agentic | 1,05 million | $10.00 | $50.00 | |
| Grok 4 | xAI | Modèle de langage volumineux (raisonnement) | 256 000 jetons | $3 | $15 | |
| Kimi K2.7 Code | Moonshot AI | LLM (codage, MoE) | 256 K | $0.6 | $2.5 | |
| Kimi K3 | Moonshot AI | LLM (architecture MoE, raisonnement, multimodal) | 1 million | $3.00 | $15.00 | |
| Kling 2.5 Turbo Pro | Kuaishou | Génération de vidéos | — | $0.07/sec | — | |
| Llama 3.1 8B | Meta | LLM (dense) | 128 K | $0.02 | $0.03 | |
| Llama 3.3 70B | Meta | LLM (dense) | 128 K | $0.10 | $0.32 | |
| Llama 4 Maverick | Meta | Multimodal (MoE) | 1 million | $0.2 | $0.8 | |
| Llama 4 Scout | Meta | Multimodal (MoE) | 10 M | $0.1 | $0.3 | |
| Mistral 7B | Mistral AI | LLM (dense) | 32 K | $0.02 | $0.03 | |
| Mistral Large 3 | Mistral AI | LLM (architecture MoE) | 256 K | $2.00 | $6.00 | |
| Mistral NeMo 12B | Mistral AI | LLM (dense) | 128 K | $0.02 | $0.04 | |
| NVIDIA Nemotron 3 Nano Omni | NVIDIA | Multimodal (omni) | 256 K | — | — | |
| Phi-4 | Microsoft | LLM (dense) | 16 K | $0.07 | $0.14 | |
| Qwen3 14B | Alibaba | LLM (dense) | 128 K | $0.12 | $0.24 | |
| Qwen3 235B-A22B | Alibaba | LLM (architecture MoE) | 128 K | $0.45 | $1.8 | |
| Qwen3 30B-A3B | Alibaba | LLM (architecture MoE) | 128 K | $0.12 | $0.5 | |
| Qwen3 32B | Alibaba | LLM (dense) | 128 K | $0.08 | $0.28 | |
| Qwen3 8B | Alibaba | LLM (dense) | 128 K | $0.04 | $0.14 | |
| Sora 2 | OpenAI | Génération de vidéos | — | $0.05/sec | — | |
| Sora 2 Pro | OpenAI | Génération de vidéos | — | $0.15/sec | — | |
| Veo 3.1 | Génération de vidéos | — | $0.05/sec | — | ||
| Wan 2.5 | Alibaba | Génération de vidéos | — | $0.05/sec | — |