GPT 6.1 Sol
A large language model in OpenAI's GPT series, designated version 6.1 with the codename suffix "Sol". It represents an incremental update within the GPT model line. Detailed technical specifications and capability profiles have not been fully disclosed publicly.
Core Facts
Timeline (last 90 days)
GPT 6.1 Sol completed the short film project in Codex using high reasoning effort, taking about 12 minutes 27 seconds from request to delivery, with a cumulative 2.755 million tokens, of which 29,884 were output and about 2.634 million were cache reads.
The finished film delivered by GPT 6.1 Sol is 54 seconds long, with clear story causality and complete music, subtitles, web page and MP4, but the stage is simple, the cat's physical performance is sparse, and the shot layering is limited.
Opus 5.5 scores 66.4% on TerminalBench, about 24 points higher than DeepSeek 1017; GPT 6.1 Sol leads 1017 slightly on DeepSWE at 75.2%.
Gemini 4 Argon's score on FrontierCode 1.1 is on par with GPT-6 Sol.
GPT 6.1 Sol was used to rewrite the entire Super Mario World from scratch for mod retro, and it has been working continuously for more than four days without finishing.
GPT 6.1 Sol demonstrated a kind of autonomous loop capability: it can judge on its own whether its output meets expectations and decide on its own what to change next, without requiring human intervention at every step.
The creator described GPT 6.1 Sol's rewriting process as "not exactly decompilation, but very close".
Anthropic first released Opus 5.5 and then Sonnet 5.5 with comparable performance at half the price, and OpenAI replicated the same strategy with GPT 6.1 Sol.
In the test, Sonnet 5.5's cache hit rate exceeded 97%, yet its cost still far surpassed that of Sol.
In the test, GPT-6.1 Sol's input tokens were 6.8 million (of which 6.6 million were cached), with output tokens of 68,000.
6 more timeline events
All Facts (11)
OpenAI officially claims that GPT-4.1 Sol is smarter than Astra in certain respects.
70%VerifiedGPT-6 Sol and Anthropic's Opus 5.5 were released on the same day.
70%UnverifiedThe finished film delivered by GPT 6.1 Sol is 54 seconds long, with clear story causality and complete music, subtitles, web page and MP4, but the stage is simple, the cat's physical performance is sparse, and the shot layering is limited.
50%UnverifiedGPT 6.1 Sol completed the short film project in Codex using high reasoning effort, taking about 12 minutes 27 seconds from request to delivery, with a cumulative 2.755 million tokens, of which 29,884 were output and about 2.634 million were cache reads.
50%UnverifiedOpus 5.5 scores 66.4% on TerminalBench, about 24 points higher than DeepSeek 1017; GPT 6.1 Sol leads 1017 slightly on DeepSWE at 75.2%.
50%UnverifiedGemini 4 Argon's score on FrontierCode 1.1 is on par with GPT-6 Sol.
50%UnverifiedGPT 6.1 Sol demonstrated a kind of autonomous loop capability: it can judge on its own whether its output meets expectations and decide on its own what to change next, without requiring human intervention at every step.
50%UnverifiedThe creator described GPT 6.1 Sol's rewriting process as "not exactly decompilation, but very close".
50%UnverifiedAnthropic first released Opus 5.5 and then Sonnet 5.5 with comparable performance at half the price, and OpenAI replicated the same strategy with GPT 6.1 Sol.
50%UnverifiedA decimal version number typically implies an incremental update rather than a major architectural overhaul.
50%UnverifiedGPT 6.1 Sol was used to rewrite the entire Super Mario World from scratch for mod retro, and it has been working continuously for more than four days without finishing.
50%