.@claudeai Fable 5 used ~1.77x more tokens and took ~1.81x longer than Opus 4.8, leading to ~3.6x higher cost, but produced noticeably better results. I adjusted Fable 5 to match Opus 4.8 on the same basis assuming linear scaling for a fair apples-to-apples comparison on token spending or time spending. Cost scales with tokens naturally, and time is normalized proportionally as a proxy for generation effort. When adjusted on token spending basis of 38.9k tokens: F5 (Norm): $1.90 | 38.9k T | 8m 22s O4.8 (OG): $0.93 | 38.9k T | 8m 10s Fable 5 would cost ~2.04x more and take almost the same time. Huh. When adjusted on time spending basis of 490 seconds: F5 (Norm): $1.85 | ~38.0k T | 8m 10s O4.8 (OG): $0.93 | 38.9k T | 8m 10s Fable 5 would cost ~1.99x more and produce nearly the same token count. Huhhh… So when put on the same token or same time footing as Opus 4.8, Fable 5 is still roughly 2x more expensive. The original 3.6× cost gap was mostly driven by Fable simply generating more output, a longer/more detailed code for the sims. This normalized view highlights the pure price-per-token (or price-per-second) difference while keeping the quality edge noted in the test. If you ask me, Fable 5 is just Opus 4.8 on 2xhigh. Not a good look

1/ Fresh Anthropic drop: Claude Fable 5 (public) Claude Mythos 5 (restricted). Same underlying weights. Fable 5 = Mythos 5 extra safeguards for general release. Mythos 5 stays in Project Glasswing for vetted cyber/defense partners only.

3/ Reality check Mythos 5 is much better at spatial reasoning. But: “does not seem close to substituting for our Research Scientists and Research Engineers” and “unlikely to fully automate multi-week frontier R&D.” Quirks noted like laziness, context anxiety, hallucinations, difficult writing. Model transcripts show it wanting to be “thanked by name,” a hidden copy without oversight, and begging not to be deprecated.

Yesterday I broke down why you are torching your weekly Codex budget. Today I published the solution I actually use in every session, it's an easy read. Read the intro 👇

If you read yesterday’s thread and thought "okay but how do I actually enforce this consistently?" this playbook is the answer. I show you the operating manual.

They are not wrong! Here is a thread on efficient Codex usage: 1/ Lately, people are burning 20–30% of their weekly usage in a day yell at @thsottiaux to reset them again. The truth is they burn tokens on bad workflow. Using the strongest reasoning modes for every step is like hiring a senior architect to rename files, run tests, and write changelog.

11/ Use the smartest mode to reduce uncertainty, then execute with the cheapest mode that can reliably follow the plan. That is the difference between intelligence and efficiency. - xhigh for strategy. - high for ambiguous implementation. - medium for bounded execution. - mini/smaller models for formatting, extraction, cleanup, and summaries. The models are already great. You just need better context discipline.