Rank of 27135from #8
Known for
Long coding and agent runs in Codex, where the bill adds up.
OpenAI's cheaper GPT-6, out on 22 September 2026 at half the price of GPT-5.6 Sol. It scores 68.8% on DeepSWE v1.1, which OpenAI puts within about a point of Claude Fable 5's best score at roughly a fifth of the cost per task.
What it does8
- 1.05M-token context, 128K output
- Six reasoning levels, none to max
- Text and image in, text out
- Knowledge cutoff April 2026
- $2 / $10 per million tokens, half of GPT-5.6 Sol
- Cached input at a tenth of the input price
- Doubles above 272K input tokens
- In ChatGPT Work and Codex, and the API as gpt-6-sol
From GPT-6 Sol’s own site. Not used for ranking.
56index score · LLMs
How well does it score?
Compared with everyone else on this board.
—votes · crowd
Reading votes…
Votes don’t change the score.