@charliermarsh: Is Tibo talking about my small-string optimization work?
Summary
Charlie Marsh asks if Tibo's praise of GPT-5.6 Sol's performance and efficiency improvements is referring to his small-string optimization work.
View Cached Full Text
Cached at: 07/29/26, 10:01 AM
Is Tibo talking about my small-string optimization work?
Tibo (@thsottiaux): Turns out GPT-5.6 Sol and friends are absolutely amazing at performance and efficiency improvements. Good time to brush up on using OpenAI models if you haven’t switched already.
Excited to be able to share more soon. Pretty wild.
Similar Articles
@omarsar0: The efficiency frontier! Where do you think GPT-5.6 will land?
Discussion of recent benchmark results for Claude Opus 4.8 and GPT-5.5 on DeepSWE Bench, with speculation about future GPT-5.6 performance and efficiency trends.
@Saboo_Shubham_: WILD times. Anthropic: Opus 5 beats GPT-5.6 on ARC-AGI-3 Tibo: I change two settings and GPT-5.6 Sol is now SoTA
Anthropic's Opus 5 beats GPT-5.6 on ARC-AGI-3, but Tibo claims GPT-5.6 Sol becomes SoTA with two setting changes involving multi-context reasoning and canonical compaction.
Am I missing something about GPT-5.5 efficiency?
A user questions the token efficiency of GPT-5.5 versus GPT-5.4 in Codex, analyzing a chart from Artificial Analysis and praising Cursor's token performance.
@theo: Updated run just finished. GPT-6.1 Sol still crushes. Turns out it performs WAY better in Codex than in mini-swe (what …
Theo (t3.gg) reports that GPT-6.1 Sol performs significantly better in Codex than in mini-swe benchmarks used by Artificial Analysis, achieving Terminal Bench 4 scores better than Opus 5.5 at roughly 1/30th the price.
@Jason_Young1231: Tibo says that Astra and 6.1sol have seen a 50% speed boost. I tested it out, and uh... 20tps -> 30tps is indeed +50%
Tibo announces a ~50% speed optimization for GPT-6 Astra and GPT-6.1 Sol across OpenAI products and third-party partners using Sign in With ChatGPT, and a user's test confirms a jump from ~20 tps to ~30 tps.