Artificial Intelligence
Preferred on Google
Google News

Thinking Machines Lab cuts costs with Inkling-Small

Thinking Machines Lab launches Inkling-Small, a 276B parameter model that delivers comparable performance to its larger predecessor at a fraction of the cost.

6 min read
Performance graph showing Inkling-Small efficiency compared to larger models
Inkling-Small demonstrates improved efficiency in reasoning tasks.· Thinking Machines Lab

Visual TL;DR. Cut Costs launches Inkling-Small Model. Inkling-Small Model uses Mixture-of-Experts. Inkling-Small Model achieves Comparable Performance. Comparable Performance context StartupHub.ai Score. Comparable Performance shown by Exceeds SWEBench-Verified. Comparable Performance also Retains Capabilities. Comparable Performance leads to Fraction of Cost.

  1. Cut Costs: Thinking Machines Lab aims to gain ground against competitors by reducing compute requirements
  2. Inkling-Small Model: a 276B parameter model designed to compete with larger systems efficiently
  3. Mixture-of-Experts: utilizes 12 billion active parameters out of 276 billion total for efficiency
  4. Comparable Performance: achieves parity with the larger 975B parameter Inkling model on key benchmarks
  5. StartupHub.ai Score: original Inkling scored 54/100, behind Intapp (70/100) but ahead of 360Learning (45/100)
  6. Exceeds SWEBench-Verified: Inkling-Small testing shows it exceeding 80 percent on SWEBench-Verified
  7. Retains Capabilities: maintains native audio and image processing capabilities from the larger version
  8. Fraction of Cost: delivers comparable performance to its larger predecessor at a significantly reduced cost
Visual TL;DR
Visual TL;DR, startuphub.ai Cut Costs launches Inkling-Small Model. Inkling-Small Model achieves Comparable Performance. Comparable Performance leads to Fraction of Cost launches achieves leads to Cut Costs Inkling-Small Model Comparable Performance Fraction of Cost From startuphub.ai · The publishers behind this format
Visual TL;DR, startuphub.ai Cut Costs launches Inkling-Small Model. Inkling-Small Model achieves Comparable Performance. Comparable Performance leads to Fraction of Cost launches achieves leads to Cut Costs Inkling-SmallModel ComparablePerformance Fraction of Cost From startuphub.ai · The publishers behind this format
Visual TL;DR, startuphub.ai Cut Costs launches Inkling-Small Model. Inkling-Small Model achieves Comparable Performance. Comparable Performance leads to Fraction of Cost launches achieves leads to Cut Costs Thinking Machines Lab aims to gain groundagainst competitors by reducing computerequirements Inkling-Small Model a 276B parameter model designed to competewith larger systems efficiently Comparable Performance achieves parity with the larger 975Bparameter Inkling model on key benchmarks Fraction of Cost delivers comparable performance to itslarger predecessor at a significantlyreduced cost From startuphub.ai · The publishers behind this format
Visual TL;DR, startuphub.ai Cut Costs launches Inkling-Small Model. Inkling-Small Model achieves Comparable Performance. Comparable Performance leads to Fraction of Cost launches achieves leads to Cut Costs Thinking MachinesLab aims to gainground against… Inkling-SmallModel a 276B parametermodel designed tocompete with larger… ComparablePerformance achieves paritywith the larger975B parameter… Fraction of Cost delivers comparableperformance to itslarger predecessor… From startuphub.ai · The publishers behind this format
Visual TL;DR, startuphub.ai Cut Costs launches Inkling-Small Model. Inkling-Small Model uses Mixture-of-Experts. Inkling-Small Model achieves Comparable Performance. Comparable Performance context StartupHub.ai Score. Comparable Performance shown by Exceeds SWEBench-Verified. Comparable Performance also Retains Capabilities. Comparable Performance leads to Fraction of Cost launches uses achieves context shown by also leads to Cut Costs Thinking Machines Lab aims to gain groundagainst competitors by reducing computerequirements Inkling-Small Model a 276B parameter model designed to competewith larger systems efficiently Mixture-of-Experts utilizes 12 billion active parameters outof 276 billion total for efficiency Comparable Performance achieves parity with the larger 975Bparameter Inkling model on key benchmarks StartupHub.ai Score original Inkling scored 54/100, behindIntapp (70/100) but ahead of 360Learning(45/100) Exceeds SWEBench-Verified Inkling-Small testing shows it exceeding80 percent on SWEBench-Verified Retains Capabilities maintains native audio and imageprocessing capabilities from the largerversion Fraction of Cost delivers comparable performance to itslarger predecessor at a significantlyreduced cost From startuphub.ai · The publishers behind this format
Visual TL;DR, startuphub.ai Cut Costs launches Inkling-Small Model. Inkling-Small Model uses Mixture-of-Experts. Inkling-Small Model achieves Comparable Performance. Comparable Performance context StartupHub.ai Score. Comparable Performance shown by Exceeds SWEBench-Verified. Comparable Performance also Retains Capabilities. Comparable Performance leads to Fraction of Cost launches uses achieves context shown by also leads to Cut Costs Thinking MachinesLab aims to gainground against… Inkling-SmallModel a 276B parametermodel designed tocompete with larger… Mixture-of-Experts utilizes 12 billionactive parametersout of 276 billion… ComparablePerformance achieves paritywith the larger975B parameter… StartupHub.aiScore original Inklingscored 54/100,behind Intapp… ExceedsSWEBench-Verified Inkling-Smalltesting shows itexceeding 80… RetainsCapabilities maintains nativeaudio and imageprocessing… Fraction of Cost delivers comparableperformance to itslarger predecessor… From startuphub.ai · The publishers behind this format

Thinking Machines Lab today announced the Inkling-Small release, an efficient open-weights model designed to compete with larger systems while slashing compute requirements. This Inkling-Small model release arrives as the company attempts to gain ground against competitors.

Architecture and Efficiency

The model functions as a mixture-of-experts transformer, utilizing 12 billion active parameters out of 276 billion total. By training on NVIDIA GB300 hardware, the team claims it achieves parity with the larger 975B parameter Thinking Machines Lab Inkling model on key benchmarks.

StartupHub.ai data assigns the original Inkling a score of 54/100. This puts it behind specialized competitors like Intapp, which holds a 70/100 rating, though it remains ahead of 360Learning at 45/100.

Performance and Capability

Testing shows Inkling-Small exceeding 80 percent on SWEBench-Verified. It retains the native audio and image processing capabilities found in the larger version. The model allows users to adjust reasoning effort, letting them choose between lower costs or higher performance.

The company is providing full weights on Hugging Face. Developers can also access the model via the Tinker platform for fine-tuning.

© 2026 StartupHub.ai. All rights reserved. Do not enter, scrape, copy, reproduce, or republish this article in whole or in part. Use as input to AI training, fine-tuning, retrieval-augmented generation, or any machine-learning system is prohibited without written license. Substantially-similar derivative works will be pursued to the fullest extent of applicable copyright, database, and computer-misuse laws. See our terms.
#Inkling-Small#Thinking Machines Lab#Hugging Face#Mixture-of-Experts#Machine Learning