Back to timeline

Thu, July 1617:55Model/APIModel releasesAPI & pricingChinese modelsModel releases guide

Ex-OpenAI CTO Murati’s Thinking Machines Launches 975B Open-Weight Model Inkling

Decision Brief

What changedThinking Machines Lab releases Inkling, a 975-billion-parameter multimodal open-weight model that leads US open-source models on the Artificial Analysis Intelligence Index but trails behind top Chinese open-source models on some tasks.
Why it mattersWith 975B parameters approaching frontier closed-source models, its $1.87/M input tokens pricing and positioning as a fine-tuning base warrant careful cost-performance evaluation.
Who should careTeams building on model APIs
Affected stackOpenAIQwenLlama
Builder actionWorth evaluating: can OpenAI replace or complement your current model
Source confidenceMedium · Reliable media or first-hand reporting

Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, recently released Inkling, a 975-billion-parameter multimodal open-weight model. On the Artificial Analysis Intelligence Index, Inkling outperforms other US open-source models, though top Chinese open-source models still excel in certain tasks. Pricing is $1.87 per million input tokens, and the company positions it as a base for fine-tuning rather than a general-purpose powerhouse. For teams fine-tuning open-source models, Inkling offers a new option, especially for those needing large parameter counts while retaining customization. However, its input cost is higher than comparable models like Llama 3 405B (~$1.0/M tokens), and output pricing is undisclosed. Moreover, it hasn't fully surpassed Chinese open-source models (e.g., Qwen 2.5 72B is better on some tasks), so teams must assess specific task requirements.

Summary basis: official / RSS sourceCompiled from the source scope noted above; the original remains authoritative.

Sources

Related intel

留言

登入后即可留言,和其他 builder 交换实测心得。

还没有留言,抢头香。