The biggest surprise of 2026 is that the capability gap between the best open-weight/source models and the top closed models has narrowed much faster than the pricing gap.
For a company consuming 1 billion input tokens and 1 billion output tokens per month:
GPT-5.5 Pro: ~$105,000
Claude Opus 4.8: ~$30,000
DeepSeek V4 Pro: ~$5,220
DeepSeek R1: ~$2,740
When building a company today, the economic frontier looks roughly like:
DeepSeek V4 Pro / R1 for high-volume inference.
Claude Opus for premium agent workflows where reliability matters.
Most CEOs have no idea that, instead of this nuanced approach, teams are often defaulting to the most expensive models and burning through massive budgets with zero governance, audit ability and control
As control planes such as Software Factory become more standard, run-rate revenue growth for frontier labs should fall meaningfully while revenues for open models surge — because the nuanced, model-agnostic approach can be implemented to focus on customer intent, model tasking and cost management
Quite a week for open-source AI
Especially American open-source
Nemotron 3 Ultra is the most important release in quite some time
And some really cool RL and fine-tuning work from Harvey
The biggest surprise of 2026 is that the capability gap between the best open-weight/source models and the top closed models has narrowed much faster than the pricing gap.
The pricing gap remains enormous while the capability gap is quite narrow
What does this mean in practice?
For a company consuming 1 billion input tokens and 1 billion output tokens per month:
GPT-5.5 Pro: ~$105,000
Claude Opus 4.8: ~$30,000
DeepSeek V4 Pro: ~$5,220
DeepSeek R1: ~$2,740
Most CEOs have no idea that, instead of this nuanced approach, teams are often defaulting to the most expensive models and burning through massive budgets with zero governance, audit ability and control
As control planes such as Software Factory become more standard, run-rate revenue growth for frontier labs should fall meaningfully while revenues for open models surge — because the nuanced, model-agnostic approach can be implemented to focus on customer intent, model tasking and cost management