[
  {
    "id": "deepseek-ai/DeepSeek-V4-Flash",
    "org": "deepseek-ai",
    "name": "DeepSeek-V4-Flash",
    "category": "text-generation",
    "tier": [
      "priority",
      "flex"
    ],
    "desc_de": "Schnelles MoE (284B/13B aktiv). 1M Kontext. Standardmodell.",
    "desc_en": "Efficient MoE, fast inference, high throughput.",
    "price_in": 0.09,
    "price_out": 0.18,
    "price_cached": 0.018,
    "context": "1024k",
    "quant": "fp4",
    "prompt": "Bleib auf DeepSeek V4 Flash",
    "url": "https://deepinfra.com/deepseek-ai/DeepSeek-V4-Flash"
  },
  {
    "id": "deepseek-ai/DeepSeek-V4-Pro",
    "org": "deepseek-ai",
    "name": "DeepSeek-V4-Pro",
    "category": "text-generation",
    "tier": [
      "priority",
      "flex"
    ],
    "desc_de": "MoE (1.6T/49B aktiv). Fortschrittliches Reasoning, Coding, Agent-Tasks.",
    "desc_en": "Advanced reasoning, coding, long-running agent tasks.",
    "price_in": 1.3,
    "price_out": 2.6,
    "price_cached": 0.1,
    "context": "1024k",
    "quant": "fp4",
    "prompt": "Wechsel zu DeepSeek V4 Pro",
    "url": "https://deepinfra.com/deepseek-ai/DeepSeek-V4-Pro"
  },
  {
    "id": "deepseek-ai/DeepSeek-V3.2",
    "org": "deepseek-ai",
    "name": "DeepSeek-V3.2",
    "category": "text-generation",
    "tier": [
      "flex"
    ],
    "desc_de": "IMO/IOI 2025 Gold. Sparse Attention. Tool-Use.",
    "desc_en": "Gold medal IMO/IOI 2025, sparse attention, tool-use.",
    "price_in": 0.26,
    "price_out": 0.38,
    "price_cached": 0.13,
    "context": "160k",
    "quant": "fp4",
    "prompt": "Nutze DeepSeek V3.2",
    "url": "https://deepinfra.com/deepseek-ai/DeepSeek-V3.2"
  },
  {
    "id": "zai-org/GLM-5.2",
    "org": "zai-org",
    "name": "GLM-5.2",
    "category": "text-generation",
    "tier": [
      "priority",
      "flex"
    ],
    "desc_de": "Z-AI Flaggschiff. 1M-Kontext. Long-Horizon-Tasks.",
    "desc_en": "Z-AI flagship, 1M context, long-horizon tasks.",
    "price_in": 0.93,
    "price_out": 3.0,
    "price_cached": 0.18,
    "context": "1024k",
    "quant": "fp4",
    "prompt": "Nutze GLM-5.2 für diese langlaufende Analyse",
    "url": "https://deepinfra.com/zai-org/GLM-5.2"
  },
  {
    "id": "zai-org/GLM-5.1",
    "org": "zai-org",
    "name": "GLM-5.1",
    "category": "text-generation",
    "tier": [
      "flex"
    ],
    "desc_de": "Coding-fokussiert. SWE-Bench Pro SOTA.",
    "desc_en": "Coding agent, SWE-Bench Pro SOTA.",
    "price_in": 1.05,
    "price_out": 3.5,
    "price_cached": 0.205,
    "context": "198k",
    "quant": "fp4",
    "prompt": "Verwende GLM-5.1 für Coding/Engineering",
    "url": "https://deepinfra.com/zai-org/GLM-5.1"
  },
  {
    "id": "zai-org/GLM-5",
    "org": "zai-org",
    "name": "GLM-5",
    "category": "text-generation",
    "tier": [
      "flex"
    ],
    "desc_de": "Open-Source. Long-Context-Reasoning, Tool-Orchestrierung.",
    "desc_en": "Advanced LLM, long-context reasoning, multi-step tools.",
    "price_in": 0.6,
    "price_out": 2.08,
    "price_cached": 0.12,
    "context": "198k",
    "quant": "fp4",
    "prompt": "Nutze GLM-5 für komplexes Reasoning",
    "url": "https://deepinfra.com/zai-org/GLM-5"
  },
  {
    "id": "zai-org/GLM-4.7-Flash",
    "org": "zai-org",
    "name": "GLM-4.7-Flash",
    "category": "text-generation",
    "tier": [
      "flex"
    ],
    "desc_de": "30B-A3B MoE. Stärkstes der 30B-Klasse. Leicht & schnell.",
    "desc_en": "Best-in-class 30B MoE, lightweight deployment.",
    "price_in": 0.06,
    "price_out": 0.4,
    "price_cached": 0.01,
    "context": "198k",
    "quant": "bf16",
    "prompt": "Nutze GLM-4.7-Flash (schnell & günstig)",
    "url": "https://deepinfra.com/zai-org/GLM-4.7-Flash"
  },
  {
    "id": "Qwen/Qwen3.6-35B-A3B",
    "org": "Qwen",
    "name": "Qwen3.6-35B-A3B",
    "category": "text-generation",
    "tier": [
      "priority",
      "flex"
    ],
    "desc_de": "Alibaba Flaggschiff-MoE (256 Experten). 35B total, 3B aktiv.",
    "desc_en": "Alibaba flagship MoE, 256 experts, 3B active.",
    "price_in": 0.15,
    "price_out": 0.95,
    "price_cached": null,
    "context": "256k",
    "quant": "fp8",
    "prompt": "Wechsel zu Qwen3.6-35B-A3B",
    "url": "https://deepinfra.com/Qwen/Qwen3.6-35B-A3B"
  },
  {
    "id": "Qwen/Qwen3.5-397B-A17B",
    "org": "Qwen",
    "name": "Qwen3.5-397B-A17B",
    "category": "text-generation",
    "tier": [
      "priority",
      "flex"
    ],
    "desc_de": "397B/17B aktiv. Thinking-Modus, MCP, 201 Sprachen. SOTA.",
    "desc_en": "397B MoE, thinking mode, MCP tool calling, 201 languages.",
    "price_in": 0.45,
    "price_out": 3.0,
    "price_cached": 0.22,
    "context": "256k-1M",
    "quant": "fp8",
    "prompt": "Nutze Qwen3.5-397B für diese anspruchsvolle Aufgabe",
    "url": "https://deepinfra.com/Qwen/Qwen3.5-397B-A17B"
  },
  {
    "id": "Qwen/Qwen3-Max",
    "org": "Qwen",
    "name": "Qwen3-Max",
    "category": "text-generation",
    "tier": [
      "partner"
    ],
    "desc_de": "Alibabas stärkstes Modell. SOTA in Wissen, Reasoning, Coding.",
    "desc_en": "Qwen flagship, SOTA across all benchmarks.",
    "price_in": 1.2,
    "price_out": 6.0,
    "price_cached": 0.24,
    "context": "250k",
    "quant": null,
    "prompt": "Wechsel zu Qwen3-Max",
    "url": "https://deepinfra.com/Qwen/Qwen3-Max"
  },
  {
    "id": "Qwen/Qwen3-Max-Thinking",
    "org": "Qwen",
    "name": "Qwen3-Max-Thinking",
    "category": "text-generation",
    "tier": [
      "partner"
    ],
    "desc_de": "Thinking-Variante von Qwen3-Max. Test-Time-Scaling.",
    "desc_en": "Thinking variant, adaptive tool-use, test-time scaling.",
    "price_in": 1.2,
    "price_out": 6.0,
    "price_cached": 0.24,
    "context": "250k",
    "quant": null,
    "prompt": "Aktiviere Qwen3-Max-Thinking",
    "url": "https://deepinfra.com/Qwen/Qwen3-Max-Thinking"
  },
  {
    "id": "moonshotai/Kimi-K2.7-Code",
    "org": "moonshotai",
    "name": "Kimi-K2.7-Code",
    "category": "reasoning-coding",
    "tier": [
      "priority",
      "flex"
    ],
    "desc_de": "Coding-agentisches Modell. 30% weniger Thinking-Tokens.",
    "desc_en": "Coding-focused agentic model, 30% fewer thinking tokens.",
    "price_in": 0.74,
    "price_out": 3.5,
    "price_cached": 0.15,
    "context": "256k",
    "quant": "fp4",
    "prompt": "Wechsel zu Kimi K2.7 Code für diese Programmieraufgabe",
    "url": "https://deepinfra.com/moonshotai/Kimi-K2.7-Code"
  },
  {
    "id": "moonshotai/Kimi-K2.6",
    "org": "moonshotai",
    "name": "Kimi-K2.6",
    "category": "reasoning-coding",
    "tier": [
      "priority",
      "flex"
    ],
    "desc_de": "Multimodal-agentisch. Langlaufendes Coding, Swarm-Orchestrierung.",
    "desc_en": "Native multimodal agentic model, long-horizon coding.",
    "price_in": 0.75,
    "price_out": 3.5,
    "price_cached": 0.15,
    "context": "256k",
    "quant": "fp4",
    "prompt": "Nutze Kimi K2.6",
    "url": "https://deepinfra.com/moonshotai/Kimi-K2.6"
  },
  {
    "id": "moonshotai/Kimi-K2.5",
    "org": "moonshotai",
    "name": "Kimi-K2.5",
    "category": "reasoning-coding",
    "tier": [
      "priority",
      "flex"
    ],
    "desc_de": "Multimodal (Text+Vision). Instant- & Thinking-Modi.",
    "desc_en": "Multimodal agentic, vision + text, instant & thinking modes.",
    "price_in": 0.45,
    "price_out": 2.25,
    "price_cached": 0.07,
    "context": "256k",
    "quant": "fp4",
    "prompt": "Nutze Kimi K2.5",
    "url": "https://deepinfra.com/moonshotai/Kimi-K2.5"
  },
  {
    "id": "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B",
    "org": "nvidia",
    "name": "Nemotron-3-Ultra-550B-A55B",
    "category": "reasoning-coding",
    "tier": [
      "priority",
      "flex"
    ],
    "desc_de": "NVIDIA Flaggschiff. Frontier Reasoning, Deep Research.",
    "desc_en": "NVIDIA flagship, frontier reasoning, orchestration, 1M context.",
    "price_in": 0.5,
    "price_out": 2.2,
    "price_cached": 0.1,
    "context": "256k",
    "quant": "fp4",
    "prompt": "Nutze NVIDIA Nemotron 3 Ultra",
    "url": "https://deepinfra.com/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B"
  },
  {
    "id": "nvidia/NVIDIA-Nemotron-3-Super-120B-A12B",
    "org": "nvidia",
    "name": "Nemotron-3-Super-120B-A12B",
    "category": "reasoning-coding",
    "tier": [
      "priority",
      "flex"
    ],
    "desc_de": "Hybrid-MoE für Multi-Agent-Anwendungen. Hohe Effizienz.",
    "desc_en": "Hybrid MoE, multi-agent applications, single GPU.",
    "price_in": 0.085,
    "price_out": 0.4,
    "price_cached": null,
    "context": "256k",
    "quant": "bf16",
    "prompt": "Nutze NVIDIA Nemotron 3 Super",
    "url": "https://deepinfra.com/nvidia/NVIDIA-Nemotron-3-Super-120B-A12B"
  },
  {
    "id": "XiaomiMiMo/MiMo-V2.5-Pro",
    "org": "XiaomiMiMo",
    "name": "MiMo-V2.5-Pro",
    "category": "reasoning-coding",
    "tier": [
      "priority",
      "flex"
    ],
    "desc_de": "Xiaomi MoE (1.02T/42B). Hybride Attention, 1M Kontext.",
    "desc_en": "Xiaomi MoE, hybrid attention, multi-token prediction.",
    "price_in": 1.0,
    "price_out": 3.0,
    "price_cached": 0.2,
    "context": "1024k",
    "quant": "fp8",
    "prompt": "Nutze MiMo V2.5 Pro",
    "url": "https://deepinfra.com/XiaomiMiMo/MiMo-V2.5-Pro"
  },
  {
    "id": "ByteDance/Seed-2.0-Pro",
    "org": "ByteDance",
    "name": "Seed-2.0-Pro",
    "category": "reasoning-coding",
    "tier": [
      "partner"
    ],
    "desc_de": "Multi-Step Planning, visuelles Reasoning, Video.",
    "desc_en": "Multi-step planning, visual reasoning, video understanding.",
    "price_in": 0.5,
    "price_out": 3.0,
    "price_cached": 0.1,
    "context": "250k",
    "quant": null,
    "prompt": "Nutze ByteDance Seed 2.0 Pro",
    "url": "https://deepinfra.com/ByteDance/Seed-2.0-pro"
  },
  {
    "id": "ByteDance/Seed-2.0-Code",
    "org": "ByteDance",
    "name": "Seed-2.0-Code",
    "category": "reasoning-coding",
    "tier": [
      "partner"
    ],
    "desc_de": "Coding-Modell für Claude Code / IDEs. Skills-Unterstützung.",
    "desc_en": "Coding optimized for IDEs, Claude Code, Skills support.",
    "price_in": 0.5,
    "price_out": 3.0,
    "price_cached": 0.1,
    "context": "250k",
    "quant": null,
    "prompt": "Nutze ByteDance Seed 2.0 Code",
    "url": "https://deepinfra.com/ByteDance/Seed-2.0-code"
  },
  {
    "id": "ByteDance/Seed-2.0-Mini",
    "org": "ByteDance",
    "name": "Seed-2.0-Mini",
    "category": "reasoning-coding",
    "tier": [
      "partner"
    ],
    "desc_de": "Niedrige Latenz, hohe Concurrency. Vierstufiges Thinking.",
    "desc_en": "Low-latency, high-concurrency, four-tier thinking.",
    "price_in": 0.1,
    "price_out": 0.4,
    "price_cached": 0.02,
    "context": "250k",
    "quant": null,
    "prompt": "Nutze ByteDance Seed 2.0 Mini",
    "url": "https://deepinfra.com/ByteDance/Seed-2.0-mini"
  },
  {
    "id": "ByteDance/Seed-1.8",
    "org": "ByteDance",
    "name": "Seed-1.8",
    "category": "text-generation",
    "tier": [
      "partner"
    ],
    "desc_de": "Multimodaler Agent. Agent-Fähigkeiten, flexibles Kontext-Management.",
    "desc_en": "Multimodal agent, enhanced agent capabilities.",
    "price_in": 0.25,
    "price_out": 2.0,
    "price_cached": 0.05,
    "context": "250k",
    "quant": null,
    "prompt": "Nutze ByteDance Seed 1.8",
    "url": "https://deepinfra.com/ByteDance/Seed-1.8"
  },
  {
    "id": "google/gemma-4-26B-A4B-it",
    "org": "google",
    "name": "Gemma-4-26B-A4B",
    "category": "multimodal",
    "tier": [
      "priority",
      "flex"
    ],
    "desc_de": "Google DeepMind MoE. Multimodal (Text+Bild). Effizient.",
    "desc_en": "Google open multimodal MoE, text+image input.",
    "price_in": 0.07,
    "price_out": 0.34,
    "price_cached": null,
    "context": "256k",
    "quant": "fp8",
    "prompt": "Nutze Gemma 4 26B: [Frage zu Bild]",
    "url": "https://deepinfra.com/google/gemma-4-26B-A4B-it"
  },
  {
    "id": "google/gemma-4-31B-it",
    "org": "google",
    "name": "Gemma-4-31B",
    "category": "multimodal",
    "tier": [
      "priority",
      "flex"
    ],
    "desc_de": "Google DeepMind Dense. Multimodal (Text+Bild).",
    "desc_en": "Google open multimodal, text+image input.",
    "price_in": 0.13,
    "price_out": 0.38,
    "price_cached": null,
    "context": "256k",
    "quant": "fp8",
    "prompt": "Nutze Gemma 4 31B: [Frage zu Bild]",
    "url": "https://deepinfra.com/google/gemma-4-31B-it"
  },
  {
    "id": "Qwen/Qwen3-VL-235B-A22B-Instruct",
    "org": "Qwen",
    "name": "Qwen3-VL-235B-A22B",
    "category": "multimodal",
    "tier": [
      "priority",
      "flex"
    ],
    "desc_de": "Stärkstes Vision-Language. Text, Bild, Video, Agent-Fähigkeiten.",
    "desc_en": "Best vision-language, text+image+video understanding.",
    "price_in": 0.2,
    "price_out": 0.88,
    "price_cached": 0.11,
    "context": "256k",
    "quant": "fp8",
    "prompt": "Analysiere mit Qwen3-VL-235B: [Bild/URL]",
    "url": "https://deepinfra.com/Qwen/Qwen3-VL-235B-A22B-Instruct"
  },
  {
    "id": "Qwen/Qwen3-VL-30B-A3B-Instruct",
    "org": "Qwen",
    "name": "Qwen3-VL-30B-A3B",
    "category": "multimodal",
    "tier": [
      "priority",
      "flex"
    ],
    "desc_de": "Leichte MoE-VL. Günstig, leistungsfähig.",
    "desc_en": "Lightweight VL MoE, affordable vision tasks.",
    "price_in": 0.15,
    "price_out": 0.6,
    "price_cached": null,
    "context": "256k",
    "quant": "fp8",
    "prompt": "Analysiere mit Qwen3-VL-30B: [Bild]",
    "url": "https://deepinfra.com/Qwen/Qwen3-VL-30B-A3B-Instruct"
  },
  {
    "id": "google/gemini-2.5-pro",
    "org": "google",
    "name": "Gemini-2.5-Pro",
    "category": "multimodal",
    "tier": [
      "partner"
    ],
    "desc_de": "Googles Thinking-Modell. Tiefstes Reasoning, 976k Kontext.",
    "desc_en": "Google thinking model, deep reasoning, 976k context.",
    "price_in": 2.5,
    "price_out": 10.0,
    "price_cached": null,
    "context": "976k",
    "quant": null,
    "prompt": "Nutze Gemini 2.5 Pro: [Frage]",
    "url": "https://deepinfra.com/google/gemini-2.5-pro"
  },
  {
    "id": "google/gemini-2.5-flash",
    "org": "google",
    "name": "Gemini-2.5-Flash",
    "category": "multimodal",
    "tier": [
      "partner"
    ],
    "desc_de": "Schnelles Thinking-Modell. Günstiger als Pro.",
    "desc_en": "Fast thinking model, lower cost.",
    "price_in": 0.3,
    "price_out": 2.5,
    "price_cached": null,
    "context": "976k",
    "quant": null,
    "prompt": "Nutze Gemini 2.5 Flash: [Frage]",
    "url": "https://deepinfra.com/google/gemini-2.5-flash"
  },
  {
    "id": "black-forest-labs/FLUX-2-klein-9b",
    "org": "black-forest-labs",
    "name": "FLUX-2-klein-9b",
    "category": "text-to-image",
    "tier": [],
    "desc_de": "Beste Qualität-zu-Latenz der FLUX-2-Familie.",
    "desc_en": "Best quality-to-latency ratio, SOTA image generation.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.015 × (w/1024) × (h/1024)",
    "context": null,
    "quant": null,
    "prompt": "Generiere ein Bild via DeepInfra: FLUX 2 klein 9b, [Prompt]",
    "url": "https://deepinfra.com/black-forest-labs/FLUX-2-klein-9b"
  },
  {
    "id": "black-forest-labs/FLUX-2-klein-4b",
    "org": "black-forest-labs",
    "name": "FLUX-2-klein-4b",
    "category": "text-to-image",
    "tier": [],
    "desc_de": "Schnellstes FLUX-2-Modell. Hoher Durchsatz.",
    "desc_en": "Fastest FLUX-2 model, high throughput.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.014 × (w/1024) × (h/1024)",
    "context": null,
    "quant": null,
    "prompt": "DeepInfra: FLUX 2 klein 4b, [Prompt]",
    "url": "https://deepinfra.com/black-forest-labs/FLUX-2-klein-4b"
  },
  {
    "id": "ByteDance/Seedream-4.5",
    "org": "ByteDance",
    "name": "Seedream-4.5",
    "category": "text-to-image",
    "tier": [
      "partner"
    ],
    "desc_de": "ByteDance Bildgen. Bessere Edit-Konsistenz, Gesichter, Text.",
    "desc_en": "ByteDance image gen, better editing, faces, text.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.04 / Bild",
    "context": null,
    "quant": null,
    "prompt": "⚠️ Captcha nötig – nutze FLUX-2-klein-9b stattdessen",
    "url": "https://deepinfra.com/ByteDance/Seedream-4.5"
  },
  {
    "id": "ByteDance/Seedream-4",
    "org": "ByteDance",
    "name": "Seedream-4",
    "category": "text-to-image",
    "tier": [
      "partner"
    ],
    "desc_de": "Multimodal: Text+Bild+Video Input. Multi-Image-Blending.",
    "desc_en": "Multimodal image model, text+image+video inputs.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.04 / Bild",
    "context": null,
    "quant": null,
    "prompt": "⚠️ Captcha nötig",
    "url": "https://deepinfra.com/ByteDance/Seedream-4"
  },
  {
    "id": "Qwen/Qwen-Image-Max",
    "org": "Qwen",
    "name": "Qwen-Image-Max",
    "category": "text-to-image",
    "tier": [
      "partner"
    ],
    "desc_de": "Realistischer, weniger AI-Look. Bessere Texturen.",
    "desc_en": "More realistic, less AI-look, better textures.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.075 / Bild",
    "context": null,
    "quant": null,
    "prompt": "⚠️ Captcha nötig",
    "url": "https://deepinfra.com/Qwen/Qwen-Image-Max"
  },
  {
    "id": "Qwen/Qwen-Image-Edit",
    "org": "Qwen",
    "name": "Qwen-Image-Edit",
    "category": "text-to-image",
    "tier": [],
    "desc_de": "Bildbearbeitung: Text ändern, Style-Transfer, Viewpoint.",
    "desc_en": "Image editing: text changes, style transfer, element adjustments.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.025 × (w/1024) × (h/1024) × (iters/25)",
    "context": null,
    "quant": null,
    "prompt": "DeepInfra: Qwen-Image-Edit, [Bild], [Änderung]",
    "url": "https://deepinfra.com/Qwen/Qwen-Image-Edit"
  },
  {
    "id": "Qwen/Qwen-Image-Edit-Max",
    "org": "Qwen",
    "name": "Qwen-Image-Edit-Max",
    "category": "text-to-image",
    "tier": [
      "partner"
    ],
    "desc_de": "Erweiterte Bildbearbeitung. LoRA, Character-Consistency.",
    "desc_en": "Enhanced image editing, LoRA, character consistency.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.075 / Bild",
    "context": null,
    "quant": null,
    "prompt": "⚠️ Captcha nötig",
    "url": "https://deepinfra.com/Qwen/Qwen-Image-Edit-Max"
  },
  {
    "id": "Bria/Bria-3.2",
    "org": "Bria",
    "name": "Bria-3.2",
    "category": "text-to-image",
    "tier": [
      "partner"
    ],
    "desc_de": "Kommerziell lizenziert. 4B Parameter. Exzellente Ästhetik.",
    "desc_en": "Commercial-ready, licensed data, 4B params.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.04 / Bild",
    "context": null,
    "quant": null,
    "prompt": "⚠️ Captcha nötig",
    "url": "https://deepinfra.com/Bria/Bria-3.2"
  },
  {
    "id": "Bria/remove_background",
    "org": "Bria",
    "name": "Bria-Remove-Background",
    "category": "text-to-image",
    "tier": [
      "partner"
    ],
    "desc_de": "Hintergrund entfernen. Lizenzierte Daten.",
    "desc_en": "Background removal, licensed data.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.018 / Bild",
    "context": null,
    "quant": null,
    "prompt": "⚠️ Captcha nötig",
    "url": "https://deepinfra.com/Bria/remove_background"
  },
  {
    "id": "Bria/erase",
    "org": "Bria",
    "name": "Bria-Erase",
    "category": "text-to-image",
    "tier": [
      "partner"
    ],
    "desc_de": "Objekte aus Bildern entfernen.",
    "desc_en": "Remove objects from images.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.04 / Bild",
    "context": null,
    "quant": null,
    "prompt": "⚠️ Captcha nötig",
    "url": "https://deepinfra.com/Bria/erase"
  },
  {
    "id": "ClarityAI/creative",
    "org": "ClarityAI",
    "name": "ClarityAI-Creative",
    "category": "text-to-image",
    "tier": [
      "partner"
    ],
    "desc_de": "AI Upscaler. Details verbessern, Realismus hinzufügen.",
    "desc_en": "AI upscaler, enhance details, add realism.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.05 / Bild",
    "context": null,
    "quant": null,
    "prompt": "⚠️ Captcha nötig",
    "url": "https://deepinfra.com/ClarityAI/creative"
  },
  {
    "id": "ACE-Step/acestep-v15-xl-sft",
    "org": "ACE-Step",
    "name": "ACE-Step-v1.5",
    "category": "text-to-music",
    "tier": [],
    "desc_de": "Text zu komplettem Song (Vocals, Lyrics, Instrumente).",
    "desc_en": "Text to full song with vocals, lyrics, instrumentation.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.001 / Sekunde Audio",
    "context": null,
    "quant": null,
    "prompt": "Generiere Musik via DeepInfra: ACE Step v1.5, [Beschreibung]",
    "url": "https://deepinfra.com/ACE-Step/acestep-v15-xl-sft"
  },
  {
    "id": "Qwen/Qwen3-TTS",
    "org": "Qwen",
    "name": "Qwen3-TTS",
    "category": "text-to-speech",
    "tier": [],
    "desc_de": "9 Preset-Voices, Voice Cloning, 10 Sprachen (🇩🇪), Streaming.",
    "desc_en": "9 preset voices, voice cloning, 10 languages, streaming.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$20.00 / 1M Zeichen",
    "context": null,
    "quant": null,
    "prompt": "DeepInfra TTS: Qwen3-TTS, [Text], Stimme=Vivian, Sprache=de",
    "url": "https://deepinfra.com/Qwen/Qwen3-TTS"
  },
  {
    "id": "Qwen/Qwen3-TTS-VoiceDesign",
    "org": "Qwen",
    "name": "Qwen3-TTS-VoiceDesign",
    "category": "text-to-speech",
    "tier": [],
    "desc_de": "Stimme per Text beschreiben – Modell generiert sie.",
    "desc_en": "Describe voice in natural language, model generates it.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$20.00 / 1M Zeichen",
    "context": null,
    "quant": null,
    "prompt": "DeepInfra TTS VoiceDesign: [Stimmbeschreibung], [Text]",
    "url": "https://deepinfra.com/Qwen/Qwen3-TTS-VoiceDesign"
  },
  {
    "id": "ResembleAI/chatterbox-multilingual",
    "org": "ResembleAI",
    "name": "Chatterbox-Multilingual",
    "category": "text-to-speech",
    "tier": [
      "priority"
    ],
    "desc_de": "Open-Source TTS. 23 Sprachen (🇩🇪). MIT-Lizenz.",
    "desc_en": "Open-source TTS, 23 languages, MIT license.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$1.00 / 1M Zeichen",
    "context": null,
    "quant": null,
    "prompt": "DeepInfra TTS: chatterbox-multilingual, [Text]",
    "url": "https://deepinfra.com/ResembleAI/chatterbox-multilingual"
  },
  {
    "id": "ResembleAI/chatterbox-turbo",
    "org": "ResembleAI",
    "name": "Chatterbox-Turbo",
    "category": "text-to-speech",
    "tier": [
      "priority"
    ],
    "desc_de": "350M Parameter. Extrem schnell. Paralinguistic Tags.",
    "desc_en": "350M params, ultra-fast, paralinguistic tags.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$1.00 / 1M Zeichen",
    "context": null,
    "quant": null,
    "prompt": "DeepInfra TTS: chatterbox-turbo, [Text]",
    "url": "https://deepinfra.com/ResembleAI/chatterbox-turbo"
  },
  {
    "id": "hexgrad/Kokoro-82M",
    "org": "hexgrad",
    "name": "Kokoro-82M",
    "category": "text-to-speech",
    "tier": [
      "priority"
    ],
    "desc_de": "82M Parameter. Winzig, schnell, günstig. Apache-Lizenz.",
    "desc_en": "82M params, tiny, fast, cheap, Apache license.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.62 / 1M Zeichen",
    "context": null,
    "quant": null,
    "prompt": "DeepInfra TTS: Kokoro-82M, [Text]",
    "url": "https://deepinfra.com/hexgrad/Kokoro-82M"
  },
  {
    "id": "XiaomiMiMo/MiMo-V2.5-tts",
    "org": "XiaomiMiMo",
    "name": "MiMo-V2.5-TTS",
    "category": "text-to-speech",
    "tier": [
      "partner"
    ],
    "desc_de": "Natürliche Sprachausgabe. Kostenlos!",
    "desc_en": "Natural speech, free!",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "Kostenlos",
    "context": null,
    "quant": null,
    "prompt": "DeepInfra TTS: MiMo-V2.5-tts, [Text]",
    "url": "https://deepinfra.com/XiaomiMiMo/MiMo-V2.5-tts"
  },
  {
    "id": "XiaomiMiMo/MiMo-V2.5-tts-voiceclone",
    "org": "XiaomiMiMo",
    "name": "MiMo-V2.5-VoiceClone",
    "category": "text-to-speech",
    "tier": [
      "partner"
    ],
    "desc_de": "Stimmen klonen. Kostenlos!",
    "desc_en": "Voice cloning, free!",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "Kostenlos",
    "context": null,
    "quant": null,
    "prompt": "⚠️ Captcha nötig",
    "url": "https://deepinfra.com/XiaomiMiMo/MiMo-V2.5-tts-voiceclone"
  },
  {
    "id": "openai/whisper-large-v3",
    "org": "openai",
    "name": "Whisper-Large-V3",
    "category": "automatic-speech-recognition",
    "tier": [],
    "desc_de": "Spracherkennung & Übersetzung. Höchste Genauigkeit.",
    "desc_en": "Speech recognition & translation, highest accuracy.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.00045 / Minute",
    "context": null,
    "quant": null,
    "prompt": "DeepInfra ASR: whisper-large-v3, [Audio]",
    "url": "https://deepinfra.com/openai/whisper-large-v3"
  },
  {
    "id": "openai/whisper-large-v3-turbo",
    "org": "openai",
    "name": "Whisper-Large-V3-Turbo",
    "category": "automatic-speech-recognition",
    "tier": [],
    "desc_de": "8x schneller als V3. Minimale Qualitätseinbußen.",
    "desc_en": "8x faster than V3, minimal quality loss.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.00020 / Minute",
    "context": null,
    "quant": null,
    "prompt": "DeepInfra ASR: whisper-large-v3-turbo, [Audio]",
    "url": "https://deepinfra.com/openai/whisper-large-v3-turbo"
  },
  {
    "id": "Qwen/Qwen3-ASR-1.7B",
    "org": "Qwen",
    "name": "Qwen3-ASR-1.7B",
    "category": "automatic-speech-recognition",
    "tier": [],
    "desc_de": "SOTA Open-Source ASR. 30 Sprachen, 22 Dialekte.",
    "desc_en": "SOTA open-source ASR, 30 languages, 22 Chinese dialects.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.00045 / Minute",
    "context": null,
    "quant": null,
    "prompt": "DeepInfra ASR: Qwen3-ASR-1.7B, [Audio]",
    "url": "https://deepinfra.com/Qwen/Qwen3-ASR-1.7B"
  },
  {
    "id": "Qwen/Qwen3-ASR-0.6B",
    "org": "Qwen",
    "name": "Qwen3-ASR-0.6B",
    "category": "automatic-speech-recognition",
    "tier": [],
    "desc_de": "Kompakte ASR. Extrem günstig ($0.00020/Min).",
    "desc_en": "Compact ASR, extremely cheap.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.00020 / Minute",
    "context": null,
    "quant": null,
    "prompt": "DeepInfra ASR: Qwen3-ASR-0.6B, [Audio]",
    "url": "https://deepinfra.com/Qwen/Qwen3-ASR-0.6B"
  },
  {
    "id": "mistralai/Voxtral-Mini-3B-2507",
    "org": "mistralai",
    "name": "Voxtral-Mini-3B",
    "category": "automatic-speech-recognition",
    "tier": [],
    "desc_de": "Mistral Audio-Verständnis. Transkription & Übersetzung.",
    "desc_en": "Mistral audio understanding, transcription & translation.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.00100 / Minute",
    "context": "32k",
    "quant": "bf16",
    "prompt": "DeepInfra ASR: Voxtral-Mini, [Audio]",
    "url": "https://deepinfra.com/mistralai/Voxtral-Mini-3B-2507"
  },
  {
    "id": "nvidia/Nemotron-3.5-ASR-Streaming-Multilingual-0.6b",
    "org": "nvidia",
    "name": "Nemotron-3.5-ASR",
    "category": "automatic-speech-recognition",
    "tier": [],
    "desc_de": "Streaming ASR. 40+ Sprachen. Echtzeit-Untertitel.",
    "desc_en": "Streaming ASR, 40+ languages, real-time captioning.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.00020 / Minute",
    "context": null,
    "quant": null,
    "prompt": "DeepInfra ASR: Nemotron-3.5-ASR, [Audio]",
    "url": "https://deepinfra.com/nvidia/Nemotron-3.5-ASR-Streaming-Multilingual-0.6b"
  },
  {
    "id": "google/veo-3.1",
    "org": "google",
    "name": "Veo-3.1",
    "category": "text-to-video",
    "tier": [
      "partner"
    ],
    "desc_de": "Google Text-to-Video. Hochauflösend, synchronisierter Audio.",
    "desc_en": "Google text-to-video, cinematic, synced audio.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.40 / Sekunde",
    "context": null,
    "quant": null,
    "prompt": "⚠️ Captcha nötig",
    "url": "https://deepinfra.com/google/veo-3.1"
  },
  {
    "id": "google/veo-3.1-fast",
    "org": "google",
    "name": "Veo-3.1-Fast",
    "category": "text-to-video",
    "tier": [
      "partner"
    ],
    "desc_de": "Schnellere Veo-Variante. Günstiger.",
    "desc_en": "Faster Veo variant, lower cost.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.15 / Sekunde",
    "context": null,
    "quant": null,
    "prompt": "⚠️ Captcha nötig",
    "url": "https://deepinfra.com/google/veo-3.1-fast"
  },
  {
    "id": "Pixverse/Pixverse-6-T2V",
    "org": "Pixverse",
    "name": "PixVerse-V6-T2V",
    "category": "text-to-video",
    "tier": [
      "partner"
    ],
    "desc_de": "15s @ 1080p. Multi-Shot. Narrative Produktion.",
    "desc_en": "15s @ 1080p, multi-shot engine, narrative production.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.045 / Sekunde",
    "context": null,
    "quant": null,
    "prompt": "⚠️ Captcha nötig",
    "url": "https://deepinfra.com/Pixverse/Pixverse-6-T2V"
  },
  {
    "id": "ByteDance/Seedance-2.0",
    "org": "ByteDance",
    "name": "Seedance-2.0",
    "category": "text-to-video",
    "tier": [
      "partner"
    ],
    "desc_de": "ByteDance Video. Multimodale Referenzen (Bild/Video/Audio).",
    "desc_en": "ByteDance video gen, multimodal references.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$4.70 / 1M Tokens",
    "context": null,
    "quant": null,
    "prompt": "⚠️ Captcha nötig",
    "url": "https://deepinfra.com/ByteDance/Seedance-2.0"
  },
  {
    "id": "Wan-AI/Wan2.2-T2V-A14B",
    "org": "Wan-AI",
    "name": "Wan2.2-T2V-14B",
    "category": "text-to-video",
    "tier": [],
    "desc_de": "14B Video-Foundation. 2 oder 5s Clips @ 16fps.",
    "desc_en": "14B video foundation, 2 or 5 sec clips @ 16fps.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.036 / Sekunde",
    "context": null,
    "quant": null,
    "prompt": "DeepInfra: Wan2.2 T2V, [Prompt]",
    "url": "https://deepinfra.com/Wan-AI/Wan2.2-T2V-A14B"
  },
  {
    "id": "Qwen/Qwen3-Embedding-8B",
    "org": "Qwen",
    "name": "Qwen3-Embedding-8B",
    "category": "embeddings",
    "tier": [
      "priority"
    ],
    "desc_de": "Text-Embedding. 32k Kontext. 0.6B-8B Varianten.",
    "desc_en": "Text embedding, 32k context, 3 sizes.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.01 / 1M Tokens",
    "context": "32k",
    "quant": null,
    "prompt": "Nutze Qwen3-Embedding-8B für die Suche",
    "url": "https://deepinfra.com/Qwen/Qwen3-Embedding-8B"
  },
  {
    "id": "nvidia/Nemotron-3-Embed-8B",
    "org": "nvidia",
    "name": "Nemotron-3-Embed-8B",
    "category": "embeddings",
    "tier": [],
    "desc_de": "Multilingual (34 Sprachen). SOTA auf RTEB. 4096-dim.",
    "desc_en": "Multilingual embedding, 34 languages, SOTA RTEB.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.035 / 1M Tokens",
    "context": "32k",
    "quant": "fp16",
    "prompt": "Verwende Nemotron-3-Embed-8B",
    "url": "https://deepinfra.com/nvidia/Nemotron-3-Embed-8B"
  },
  {
    "id": "nvidia/Nemotron-3-Embed-1B-NVFP4",
    "org": "nvidia",
    "name": "Nemotron-3-Embed-1B",
    "category": "embeddings",
    "tier": [],
    "desc_de": "Kompakt. 2048-dim. Günstig. NVFP4-quantisiert.",
    "desc_en": "Compact embedding, 2048-dim, NVFP4 quantized.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.01 / 1M Tokens",
    "context": "32k",
    "quant": "fp4",
    "prompt": "Verwende Nemotron-3-Embed-1B",
    "url": "https://deepinfra.com/nvidia/Nemotron-3-Embed-1B-NVFP4"
  },
  {
    "id": "BAAI/bge-m3",
    "org": "BAAI",
    "name": "BGE-M3",
    "category": "embeddings",
    "tier": [
      "priority"
    ],
    "desc_de": "Multilingual (100+ Sprachen). Dense + Sparse + Multi-Vektor.",
    "desc_en": "100+ languages, dense + sparse + multi-vector.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.01 / 1M Tokens",
    "context": "8k",
    "quant": "fp32",
    "prompt": "Nutze BGE-M3 für multilinguale Suche",
    "url": "https://deepinfra.com/BAAI/bge-m3"
  },
  {
    "id": "Qwen/Qwen3-Reranker-8B",
    "org": "Qwen",
    "name": "Qwen3-Reranker-8B",
    "category": "reranker",
    "tier": [
      "priority"
    ],
    "desc_de": "Reranking: Ergebnisse nach Relevanz sortieren.",
    "desc_en": "Reranking: sort results by relevance.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.05 / 1M Tokens",
    "context": "32k",
    "quant": null,
    "prompt": "Nutze Qwen3-Reranker-8B",
    "url": "https://deepinfra.com/Qwen/Qwen3-Reranker-8B"
  },
  {
    "id": "nvidia/llama-nemotron-rerank-vl-1b-v2",
    "org": "nvidia",
    "name": "Nemotron-Reranker-VL-1B",
    "category": "reranker",
    "tier": [
      "priority"
    ],
    "desc_de": "Multimodaler Reranker: Chart, Tabellen, Infografiken.",
    "desc_en": "Multimodal reranker, charts, tables, infographics.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.01 / 1M Tokens",
    "context": "10k",
    "quant": "bf16",
    "prompt": "Nutze Nemotron-Reranker-VL",
    "url": "https://deepinfra.com/nvidia/llama-nemotron-rerank-vl-1b-v2"
  },
  {
    "id": "openai/clip-vit-base-patch32",
    "org": "openai",
    "name": "CLIP-ViT-Base-32",
    "category": "zero-shot-image-classification",
    "tier": [],
    "desc_de": "Bildklassifikation ohne Training. Vision Transformer.",
    "desc_en": "Zero-shot image classification, Vision Transformer.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.0005 / second",
    "context": null,
    "quant": null,
    "prompt": "Klassifiziere mit CLIP: [Bild]",
    "url": "https://deepinfra.com/openai/clip-vit-base-patch32"
  },
  {
    "id": "openai/clip-vit-large-patch14-336",
    "org": "openai",
    "name": "CLIP-ViT-Large-336",
    "category": "zero-shot-image-classification",
    "tier": [],
    "desc_de": "Größere CLIP-Variante. Höhere Genauigkeit.",
    "desc_en": "Larger CLIP variant, higher accuracy.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.0005 / second",
    "context": null,
    "quant": null,
    "prompt": "Klassifiziere mit CLIP-Large: [Bild]",
    "url": "https://deepinfra.com/openai/clip-vit-large-patch14-336"
  },
  {
    "id": "nvidia/Cosmos3-Nano",
    "org": "nvidia",
    "name": "Cosmos3-Nano",
    "category": "world-model",
    "tier": [],
    "desc_de": "NVIDIA World Foundation Model. Reasoner + Generator.",
    "desc_en": "NVIDIA world foundation, Reasoner + Generator towers.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.025 / second (720p)",
    "context": null,
    "quant": null,
    "prompt": "Nutze Cosmos3-Nano",
    "url": "https://deepinfra.com/nvidia/Cosmos3-Nano"
  },
  {
    "id": "nvidia/Cosmos3-Super",
    "org": "nvidia",
    "name": "Cosmos3-Super",
    "category": "world-model",
    "tier": [],
    "desc_de": "Größeres World Model. Höhere Qualität.",
    "desc_en": "Larger world model, higher quality.",
    "price_in": null,
    "price_out": null,
    "price_cached": null,
    "price_label": "$0.05 / second (720p)",
    "context": null,
    "quant": null,
    "prompt": "Nutze Cosmos3-Super",
    "url": "https://deepinfra.com/nvidia/Cosmos3-Super"
  }
]