NagaAI

Access the world's best AI models from OpenAI, Anthropic, Google, and more through a single, unified API

[NagaAI @NagaAIProject]

🚀Over the last few weeks, we've rolled out some massive additions to NagaAI! You can now use Gemini 3.1 Pro Preview, Claude 4.6 (Opus & Sonnet), and MiniMax-M2.5.

Also added: GLM-5, Qwen3 models, and Trinity Large Preview (free).

https://t.co/8bjTUUwpFU

[NagaAI @NagaAIProject]

@thomas_agius In theory you can already use us with Openclaw - https://t.co/1HjjB6e0t2. Try specifying a custom `
openai-completions provider.

But a prebuilt option in the providers list might appear in the future!

[NagaAI @NagaAIProject]

🚀Organizations, Alerts & New Models

👥Organizations:
Create teams with shared billing, API keys, and roles.
🔗https://t.co/b1SuCoG1tP

⚠️Balance Alerts:
Set email notifications for low credits.
🔗https://t.co/5RcXSsFyuh

🤖New Models:
- Kimi K2.5
- Flux 2 Family

[NagaAI @NagaAIProject]

Dashboard: https://t.co/Hak5RfdRbz
API Reference: https://t.co/ut8QPHzkug

[NagaAI @NagaAIProject]

🚀Quck Update:
1. API Key Expiration: Set keys to expire automatically. Great for temporary access (scripts/collaborators) without manual revocation.
2. Reasoning Effort: Added `none` (disable reasoning) and `xhigh` (max thinking).

[NagaAI @NagaAIProject]

Messages API Reference: https://t.co/bixK2qEDZe
Responses API Reference: https://t.co/oaykT9jzKs

[NagaAI @NagaAIProject]

🚀 New: Anthropic Messages & OpenAI Responses API Support!

We've added Messages/Responses API support, allowing you to use apps like **Claude Code** directly with NagaAI. Just change your Base URL.

Integration guides: https://t.co/X0gTh6o7Dd

[NagaAI @NagaAIProject]

🎉The Christmas promotion has been extended until January 7 at users request. Now you can get even more bonus credits!

Happy New Year!

[NagaAI @NagaAIProject]

🎄Christmas Special: Get 2x credits on any top-up of $25+!

That's double the tokens for Gemini 3, Claude 4.5, GPT 5.2, GLM, and any other models. Valid until Jan 2.

Examples:
- $25 → 50 credits
- $100 → 200 credits
- $250 → 500 credits

https://t.co/9HUGS0IdNt

[NagaAI @NagaAIProject]

- Seedream 4.5: Finally renders small text & logos correctly.
- Qwen-Edit: Native LoRA support & geometric consistency.

[NagaAI @NagaAIProject]

🚀GLM-4.7, MiniMax-M2.1, Seedream 4.5, and Qwen-Image-Edit-2511 are now live!

- GLM-4.7: https://t.co/OJsR5J00a0's flagship with robust reasoning for agents.
- MiniMax-M2.1: Lightweight 10B model, 72.5% on SWE-Bench.

[NagaAI @NagaAIProject]

Links:

https://t.co/uV7yqfV4yD
https://t.co/3seWvwP7nI

[NagaAI @NagaAIProject]

🚀 Gemini 3 Flash Preview & GPT-Image-1.5 Models are Live

- Gemini 3 Flash: High-speed, low-latency reasoning for agents. Supports 1M context, multimodal inputs, and coding workflows.
- GPT-Image-1.5: 4x faster generation with precise editing and advanced text rendering.

[NagaAI @NagaAIProject]

🚀 New Model Drop: The entire GPT-5.2 family is live!

- GPT-5.2 Pro: The most advanced model with "think hard" routing & agentic coding mastery.
- GPT-5.2: Frontier-grade adaptive reasoning for science & math.
- GPT-5.2 Chat: Lightweight & fast

https://t.co/PvpRrTA583

[NagaAI @NagaAIProject]

🚀 DeepSeek-V3.2 & V3.2-Speciale are now live.

- V3.2: GPT-5 level performance with "Thinking in Tool-Use" for advanced agents.
- V3.2-Speciale: Rivals Gemini-3.0-Pro. 🥇 Gold medals in IMO & IOI 2025.

https://t.co/Qtu04EneZ3
https://t.co/g8s6YPsv27

[NagaAI @NagaAIProject]

🚀Claude Opus 4.5 is live on NagaAI.

Anthropic's new flagship delivers SOTA performance for software engineering and agents:
- 80.9% on SWE-bench Verified (beating GPT-5.1)
- 59.3% on Terminal-bench 2.0
- 37.6% on ARC-AGI-2 (Reasoning)

https://t.co/abDZMQKhCP

[NagaAI @NagaAIProject]

🎁Launching the NagaAI Referral Program!

Share your link and earn credits:
- Give friends a 25% bonus on their first top-up.
- Earn up to 15% commission on every payment they make.

https://t.co/eZK0rfL1Gp

[NagaAI @NagaAIProject]

🎨Nano Banana Pro is live at NagaAI

Studio-quality 4K generation, advanced physics control (lighting/focus), and SOTA text rendering. Grounded with Google Search for accuracy.

Try it: https://t.co/AN6iVeXPew

[NagaAI @NagaAIProject]

🚀 Grok 4.1 Fast now available

xAI's best tool-calling model with 2M context: state-of-the-art on Berkeley Function Calling v4, Research-Eval Reka (63.9), FRAMES (87.6).

Two variants: reasoning for complex tasks, non-reasoning for speed.

https://t.co/CptPMxflhJ

[NagaAI @NagaAIProject]

Minor Updates Live:

- Chat Playground now displays reasoning chains from reasoning models
- TTS API (/v1/audio/speech) supports response_format - choose mp3, opus, aac, flac, wav, or pcm
- Response metrics now visible in Playground (TTFT, Process TIme)