Together AI brings Thinking Machines Lab’s new model Inkling on day 0

Summary

Thinking Machines Lab released Inkling, and Together AI announced day-zero access through its production inference platform. The multimodal mixture-of-experts model accepts text, image, and audio inputs and generates text through a unified decoder architecture, with controllable inference effort for trading reasoning depth, token usage, and latency. Inkling has 975B total parameters, 40B active parameters per token, a 1M-token context window, query-conditioned relative attention, short causal convolutions, and shared-sink MoE routing; Together serves it with a FlashAttention-4-based kernel. Preliminary evaluations report strong results across scientific reasoning, mathematics, coding, agentic, vision, and audio benchmarks, while the model is also post-trained for forecasting and calibrated prediction. It is available on Together Serverless through OpenAI-compatible APIs without infrastructure provisioning.

Takeaways

1. 1

   Inkling accepts text, image, and audio inputs and processes them through one decoder stack, producing text outputs without separate modality-specific pipelines.

2. 2

   The model combines query-conditioned relative attention, five local-attention layers followed by one full-attention layer, four-token causal convolutions, and shared-sink mixture-of-experts routing.

3. 3

   Developers can adjust inference effort per request to balance reasoning depth, token usage, and latency, while Together Serverless handles infrastructure and GPU provisioning.

FAQ