IBM Releases Granite 4.1 8B with 131K Context Window at $0.05/M Input Tokens
IBM has released Granite 4.1 8B, an 8-billion-parameter decoder-only language model with a 131,072-token context window. The model supports 12 languages and costs $0.05 per million input tokens and $0.10 per million output tokens, available under the Apache 2.0 license.
Granite 4.1 8B — Quick Specs
IBM Releases Granite 4.1 8B with 131K Context Window at $0.05/M Input Tokens
IBM has released Granite 4.1 8B, an 8-billion-parameter decoder-only language model with a 131,072-token context window, priced at $0.05 per million input tokens and $0.10 per million output tokens.
Model Specifications
Granite 4.1 8B is a dense transformer model with 8 billion parameters, released on April 30, 2026. The model supports a context window of 131,072 tokens, positioning it in the long-context model category alongside competitors like Claude and GPT-4.
The model is distributed under the Apache 2.0 license, making it available for both commercial and research use without licensing restrictions.
Capabilities
Granite 4.1 8B targets enterprise use cases with several specific features:
- Tool calling: Implements OpenAI-compatible function calling for integration with external systems
- Code generation: Includes fill-in-the-middle support for code completion tasks
- RAG support: Designed for retrieval-augmented generation workflows
- Text processing: Handles summarization, classification, and extraction tasks
The model supports 12 languages: English, German, Spanish, French, Japanese, Portuguese, Arabic, Czech, Italian, Korean, Dutch, and Chinese.
Deployment
IBM is distributing Granite 4.1 8B through OpenRouter, which provides routing to multiple infrastructure providers. Model weights are publicly available for self-hosting.
The pricing structure of $0.05 per million input tokens and $0.10 per million output tokens places it in the mid-range pricing tier for models of this size, comparable to other 8B-parameter models from companies like Meta and Mistral.
What This Means
Granite 4.1 8B represents IBM's continued investment in open-source enterprise AI, offering a permissively licensed alternative to proprietary models. The 131K context window is notably large for an 8B-parameter model, though actual performance at that context length remains to be independently verified. The Apache 2.0 license and multi-language support make it a viable option for enterprises requiring on-premises deployment or specific regulatory compliance, particularly in the 12 supported language markets.
Related Articles
Microsoft Releases VibeVoice-ASR-BitNet: 1.58GB Speech Recognition Model Runs Real-Time on CPU, No GPU Needed
Microsoft Research released VibeVoice-ASR-BitNet, a quantized 1.58GB version of its VibeVoice-ASR speech recognition model that achieves real-time inference (RTF < 1) on as few as 3 CPU threads. The model runs 1.6-2.3x faster than Whisper.cpp on commodity x86 and ARM hardware, with a modest accuracy tradeoff.
Moonshot AI Releases Kimi K3: 2.8T-Parameter Open-Weight Model with 1M-Token Context, Now Available via Unsloth Quantiza
Moonshot AI has released Kimi K3, a 2.8-trillion-parameter open-weight mixture-of-experts model with a 1-million-token context window and native multimodal support. Unsloth has published Dynamic 2.0 quantized versions on Hugging Face, claiming improved accuracy over other quantization methods.
Moonshot AI Releases Kimi K3: Open-Weight 2.8T-Parameter Model With 1M-Token Context and Native Multimodality
Moonshot AI has released Kimi K3, an open-weight 2.8-trillion-parameter mixture-of-experts model with 104B activated parameters, a 1,048,576-token context window, and native multimodal support. The company describes it as the world's first open 3T-class model, built on a new Kimi Delta Attention architecture.
Microsoft Releases Mage-Flow: Compact 4B Image Generation and Editing Models Matching Systems 5-8x Larger
Microsoft has released Mage-Flow, a family of 4B-parameter image generation and editing models built on a shared tokenizer-transformer stack. According to Microsoft, the Turbo variants match or beat open-source systems with 5-8x more parameters while running in 4 diffusion steps.
Comments
Loading...