ZDNET’s key takeaways
- Thinking Machines has released a competitive new open model, Inkling 2 Small.
- Most competitive open models come from China.
- The release follows French lab Mistral’s Le Chonk open model, which excels at coding.
Thinking Machines Lab, a small AI startup from ex-OpenAI CTO Mira Murati, has released its latest model, Inkling 2 Small – and according to third-party tests Artificial Analysis, it is set to give Chinese models a run for their lead. It’s technically proprietary so far, but Thinking Machines said it will release its weights next week.
Also: Open Weights vs. Closed: An AI Civil War Is
Chinese companies have led the pack on open models in recent years. Models from DeepSeek and Moonshot regularly top leaderboards and rival the capabilities of much more expensive, closed models from OpenAI and Anthropic.
What Inkling 2 Small does
Thinking Machines’ latest model has 594 billion parameters and supports image inputs but only outputs text. The company noted its web and game design capabilities and that it evaluated the model’s approach to coding issues alongside the code it generated during testing.
“The result is a model that is stronger in coding and easier to use in real agent loops,” the company said.
Also: Nvidia takes another step forward for US Open-weight AI
Inkling 2 Small scored a 31 on the Artificial Analysis Intelligence Index, a composite test “to track progress toward artificial general intelligence” that combines benchmark scores across coding, science and general categories. A score of 31 is above average for models in its class and is tied with Gemini 3.8 flash. For context, Gemini 4 Carbon and Claude Opus 5.5 are the index with 60 and 59 respectively.
The open-weight race
Thinking Machines said it will release the model’s weights next week, a practice that is becoming common. Due to the increased security concerns around AI in general, AI companies spend several weeks testing the open model with cybersecurity teams and security auditors before releasing the weights, rather than leading out of the gate with the weights.
French AI lab Mistral recently released its Le Chonk model via the same hidden process – it now keeps up with OpenAI’s GPT-6.1 Terra in coding capabilities, which is rare for Western-made open models.
Also: Mistral’s Le Chonk is now a top open model for coding
“[Thinking Machines’] Plan to publish its weights would make it the highest scoring open weight model developed outside of China based on current rankings,” the artificial analysis said.
But rarely is a model the complete package.
“Inkling 2 Small is among the leading models in intelligence and reasonable prices when you compare with other models with similar price. It is also remarkably slow per task and very verbose,” said Artificial Analysis. The organization also called Inkling 2 Small’s pricing “a bit expensive” at $0.30 per one million input tokens, as the median cost for its class is $0.24.
However, the progress of the model over its predecessor stands out for Artificial Analysis reviews: Inkling 2 Small jumped 41% in automation skills and 20% in coding skills. However, most models excel in certain areas or on specific benchmarks, which makes it difficult to consider one or the other the true best coding model overall.
Also: I tried to replace the Claude code with a local AI stack which is completely free
Inkling 2 Small also made some gains in accuracy, artificial intelligence said.
“On AA-Omniscience v1.1, Inkling Small’s hallucination rate drops from 61% to 27%, while accuracy increases from 32% to 34% and attempt rate drops from 77% to 53%,” his review reported.