Inkling: This Open-Weight Model Wants To Be Fine Tuned

Highlights

• Thinking Machines released Inkling, an open-weight model with 975 billion total parameters under an Apache 2.0 license.

• Inkling processes raw audio as frequency spectrograms and images as 40 by 40 pixel patches directly in the token space without intermediate text conversion.

• On instruction following, Inkling scores 79.8, beating GLM 5.2 at 73.3, but scores 63.8 on terminal bench compared to 82.7 for agents.

• Law firm AI tool developer Harvey trained a small model on legal citations, raising its F1 score from 0.56 to 0.68 and winning or tying GPT-4.0 in 93 percent of head-to-head tests.

• Tinker supports automated self-fine-tuning using LoRa, freezing original weights while injecting lightweight smaller matrices.

Timeline

1. 00:00 - 01:21

   Model Overview and Architecture

   Inkling targets fine-tuning applications via the Tinker platform rather than competing directly with frontier general-purpose models. The creators explicitly manage expectations regarding standalone general capability, noting strong instruction following paired with weaker autonomous agent execution.

   • Thinking Machines released its first open-weight model, Inkling, with 975 billion parameters under an Apache 2.0 license.
   • Inkling scores 79.8 on instruction following compared to GLM 5.2 at 73.3, but scores 63.8 on terminal bench against 82.7 for agent workflows.
   • The architecture utilizes a mixture of experts with 256 unique experts per layer, routing each token to six experts so that 41 billion parameters operate concurrently.
2. 01:21 - 02:46

   Native Multimodal Processing

   Unlike standard architectures that rely on separate encoders or speech models to convert audio and images into text before language model ingestion, Inkling processes raw audio and image patches concurrently within the same token space. This native ingestion minimizes information degradation.

   • Raw audio enters the model as frequency spectrograms split into frequency bands without prior speech-to-text transcription.
   • Images enter the model as 40 by 40 pixel patches directly into the shared token space alongside text tokens.
   • Eliminating intermediate text transcription layers prevents data compression and loss during multimodal input ingestion.
3. 02:46 - 04:10

   Fine-Tuning Strategy and Performance

   Fine-tuning suits narrow, high-volume tasks requiring thousands of structured output examples that exceed prompt context limits. When applied to specific legal document extraction, smaller fine-tuned models outperform generalist flagships while operating at higher execution speeds.

   • Fine-tuning teaches models output format, tone, and structured style rather than adding new factual knowledge.
   • Retrieval systems consistently outperform fine-tuning for factual knowledge integration, and fine-tuning can degrade model chain-of-thought reasoning.
   • Harvey trained a legal citation model that achieved an F1 score increase from 0.56 to 0.68 and matched or beat GPT-4.0 in 93 percent of head-to-head tests.
4. 04:10 - 06:28

   Automated Fine-Tuning and Tinker Platform

   Users can prompt Tinker to automatically fine-tune Inkling for custom constraints, such as eliminating specific letters from generated responses. While generalist models excel without modification, specialized audio-driven tasks and granular thinking control position Inkling as a distinct option on the Tinker platform.

   • Tinker enables automated dataset creation and fine-tuning using LoRa by freezing original weights and injecting lightweight adaptation matrices.
   • Inkling stands out on Tinker as the sole platform model supporting raw audio input, accompanied by speech recognition and medical prescription recipes.
   • Inkling offers a continuous slider from zero to one to configure thinking effort rather than discrete low, medium, or high switches.