OpenAI drops GPT-6.1 Sol: Same brains as Astra, but five times cheaper
OpenAI decided that being rich isn't enough, so they launched GPT-6.1 Sol. It is basically the bargain-bin version of their flagship model that performs shockingly well for coding and documentation while making your wallet cry less.
GPT-6.1 Sol arrives as an upgrade to GPT-6 Sol specifically optimized for crunching massive codebases and automating document workflows. In the DeepSWE 1.1 benchmark, which stress-tests how models actually handle real-world programming, this new iteration effectively matches the performance of the flagship GPT-6 Astra despite being significantly cheaper to run.
When thrown into OSWorld 2.0 to see how well it navigates apps, the model showed a seven-point jump in reasoning capabilities compared to its predecessor. The pricing structure is the real headline here, with input tokens at $2, cached inputs at $0.10, and output tokens at $10. These rates represent a five-fold reduction in operational costs compared to the current Astra standard.
The rollout covers ChatGPT Work and Codex for Plus, Pro, Business, Enterprise, and Edu tiers, leaving the standard ChatGPT interface out in the cold for now. The math suggests OpenAI is betting heavily on keeping developers hooked by letting them burn through infinite tokens without immediately going bankrupt.
Source: OpenAI
Comments
Help shape the next version: Add context or suggest a correction. AI review can add points toward a rewrite. Reviews and updates may take time; a full meter does not guarantee a new version.