GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price
OpenAI released GPT-6.1 Sol, matching near-Astra capabilities on professional tasks at one-fifth the price with heavily discounted cached tokens.
summary is enough The announcement details expected benchmark and pricing updates, but the discussion provides critical context on market pressures and subscription tier changes.
summary
OpenAI has launched GPT-6.1 Sol, an upgraded model that approaches GPT-6 Astra intelligence for professional, coding, and agentic tasks at a fifth of the cost. The release features aggressive pricing, including cached inputs at $0.10 per million tokens, and strong benchmark gains on evaluations like DeepSWE and OSWorld. While available immediately for developers and enterprise channels, the announcement arrives alongside heated debate regarding plunging inference costs, competitive pressure from alternative labs, and OpenAI's controversial subscription tier adjustments.
key points
- Cached input pricing drops to $0.10 per million tokens, offering a 95% discount over standard input pricing.
- GPT-6.1 Sol approaches GPT-6 Astra performance on complex benchmarks like DeepSWE v1.1 and OSWorld 2.0 at roughly one-fifth to one-seventh the cost.
- The model is available immediately via the OpenAI API as gpt-6.1-sol and inside ChatGPT Work and Codex.
who should read it
AI developers and enterprise users looking to optimize inference and context-caching costs for agentic workflows.
discussion mixed
consensus
Commenters view the aggressive price cuts and lower cached token costs as a necessary response to fierce competition from low-cost frontier models.
pushback
Pushback centers on OpenAI's new high-priced subscription tier and reduced allowances on lower plans, alongside complaints about model regression.
I am yet to spend $200 on deepseek this year. Not sure what kind of usage can justify $200/month of either openai or anthropic, i'm not even talking about $500. Deepseek is faster, IMO intelligence difference is negligible and it so much cheaper that i no longer care about how much i use it. I never hit any daily/weekly quota or anything like that while working or tinkering. At this point i am OK with being 6 months behind the "frontier", purely on bang-for-buck basis and who cares which shadowy government gets my data.
It captures the growing developer shift toward ultra-cheap alternatives like DeepSeek, questioning the value proposition of expensive $200 or $500 monthly subscriptions.
Generated by an LLM from the article and thread. Check the sources before relying on it.