Supported Providers 🌐

NexusAI-Client unifies leading zero-cost tiers and high-performance commercial AI APIs under a standardized protocol.

📊 Summary Matrix

🔍 Provider Deep Dives

1. Cerebras ("cerebras")

• Key Advantage: Powered by the CS-3 Wafer-Scale engine, providing record-breaking generation speeds of 2,000+ tokens/second.
• Free Quotas: 30 Requests/min, 60k Tokens/min, 1,000,000 tokens/day on free tier.
• Protocol: OpenAI-compatible REST API.

async with AIGateway("cerebras") as client:
    res = await client.generate_text("Explain quantum entanglement in 1 sentence.")
    print(res.text)

2. Cohere ("cohere")

• Key Advantage: Native integration with Cohere's Command R+ enterprise reasoning models and Aya Vision multilingual multimodal models.
• Protocol: Cohere V2 REST API (/v2/chat).
• Vision Support: Yes, pass image paths or URLs to analyze_image().

async with AIGateway("cohere") as client:
    res = await client.generate_text("Analyze the market implications of semiconductor tariffs.")
    print(res.text)

3. DeepSeek ("deepseek")

• Key Advantage: Industry-leading ultra-low-cost pricing ($0.27 / 1M input tokens) with DeepSeek-V3 and DeepSeek-R1 reasoning.
• Budget Inspection: NexusAI-Client queries DeepSeek's live /user/balance endpoint to report exact remaining USD credits.

async with AIGateway("deepseek") as client:
    account = await client.get_account_info()
    print(account.format_summary())
    # Output: "Solde restant: $4.99 | Consommé: $0.0120"

4. Google Gemini Free ("gemini_free")

• Key Advantage: 1,000,000 token context window and Intelligent 11-Model Auto-Rotation: if a model encounters HTTP 429 quota limits, it instantly rotates across gemini-3.5-flash-lite, gemini-3.1-flash-lite, gemini-3.7-flash, gemini-3.6-flash, gemini-3.5-flash, gemini-2.5-flash-lite, gemini-2.5-flash, gemma-4-31b-it, and gemma-4-26b-a4b-it.
• Default Model: gemini-3.5-flash-lite (500 RPD, 15 RPM, 250k TPM).
• Vision Support: Multimodal rotation across 7 vision-capable Gemini models.
• Stateful Cooldown: Automatically tracks and skips rate-limited models without performance penalty.

async with AIGateway("gemini_free") as client:
    res = await client.generate_text("Summarize the top 5 papers on multi-agent LLM systems.")
    print(f"Served by [{res.model}]: {res.text}")

5. Google Gemini Pro ("gemini_pro")

• Key Advantage: High-capacity Google Cloud Vertex/Studio tier with zero rate-limit constraints, ideal as an enterprise fallback.
• Default Model: gemini-3.1-pro-preview.
• Protocol: Gemini REST API.

async with AIGateway("gemini_pro") as client:
    res = await client.generate_text("Perform in-depth architectural analysis of distributed systems.")
    print(f"Served by [{res.model}]: {res.text}")

6. Groq Cloud LPU ("groq")

• Key Advantage: Ultra-low latency inference powered by custom Tensor Streaming Processors (LPUs).
• Default Model: openai/gpt-oss-120b.
• Free Quotas: 30 Requests/min, 14,400 Requests/day.

async with AIGateway("groq") as client:
    async for chunk in client.stream_text("Write a short poem about speed."):
        print(chunk, end="", flush=True)

7. Mistral AI ("mistral")

• Key Advantage: Access to Codestral (specialized code generation) and Pixtral (vision), with developer-tier access.
• Vision Support: Supported via pixtral-12b-2409.

8. Nvidia NIM ("nvidia_free")

• Key Advantage: 1,000 free GPU credits provided via Nvidia NGC API catalog, hosting Llama 3.1, Nemotron, and Mistral models.

9. OpenRouter ("openrouter")

• Key Advantage: Catalog of 670+ AI models with dynamic token pricing and 19 free-tier endpoints (openrouter/free, google/gemini-2.0-flash-thinking-exp:free, etc.).
• Discovery: Automatically discovers pricing and free tags dynamically via list_models().

10. OrcaRouter ("orcarouter")

• Key Advantage: Multi-provider zero-margin gateway offering transparent routing and access to high-performance free models (qwen/qwen3.8-27b-free, DeepSeek, Llama).
• Vision Support: Supported via Qwen 3.8 Vision multimodal models.
• Free Models Detection: Automatically detects zero-cost models ending with -free or :free.
• Protocol: OpenAI-compatible REST API (https://api.orcarouter.ai/v1).
• Quota & Balance: Fixed minute/daily windows with prompt token limit inspection.

async with AIGateway("orcarouter") as client:
    res = await client.generate_text("Explain the benefits of zero-margin AI routing.")
    print(res.text)