2026年8月16日の Tibo の投稿

返信先 @xikhar

@xikhar Banger

返信 40 · リポスト 1 · いいね 361

返信先 @MarieAdant

@MarieAdant That you're missing a Codex sticker

返信 60 · リポスト 1 · いいね 422

Here is how to enable a 1M-token context window in Codex for GPT-5.6 Sol. Even though we have tuned the context limit in Codex to be set optimally when it comes to performance and cost, this is a common ask, so here it is documented. A larger context window lets Codex retain more code, tool output, and conversation history before summarizing older material. You need a model that supports it. And GPT-5.6 Sol, for example, has a documented 1,050,000-token window. Open ~/.codex/config.toml and add or update these settings at the top level, before any [section] headers: ``` model = "gpt-5.6-sol" model_context_window = 1000000 model_auto_compact_token_limit = 900000 ``` The first setting selects the model. The second tells Codex to use a one-million-token context budget. The third starts automatic history compaction around 900,000 tokens, leaving some headroom. Restart Codex client and start a new session after saving. To try the configuration for a single CLI session without changing your defaults: ``` codex -m gpt-5.6-sol \ -c model_context_window=1000000 \ -c model_auto_compact_token_limit=900000 ``` Have fun, but also know that we tuned the default carefully!

返信 996 · リポスト 1,153 · いいね 14,626

Sunday is for good vibes. As is any day of the week, but especially Sunday.

返信 646 · リポスト 81 · いいね 4,657

返信先 @Robertg761_

@Robertg761_ Indeed

返信 28 · リポスト 0 · いいね 182

返信先 @Ananth7e

@Ananth7e Maybe

返信 156 · リポスト 13 · いいね 2,339

返信先 @Ananth7e

@Ananth7e Maybe

返信 100 · リポスト 2 · いいね 914

返信先 @NotiDivad

@NotiDivad Yes

返信 93 · リポスト 7 · いいね 1,378

What's your /goal for the night?

返信 2,468 · リポスト 45 · いいね 2,815

返信先 @Ethan_Smartsys

@Ethan_Smartsys Happens to the best

返信 25 · リポスト 1 · いいね 156

返信先 @butter86396

@butter86396 Sorry, only 11pm over here.

返信 13 · リポスト 0 · いいね 408

On tokens and prices per token. I said I’d write more about this, so here goes: an OpenAI token != another model’s token. We compare AI prices in dollars per million tokens as if a token were a standardized unit, like a gram or a kilowatt-hour. It isn’t. Different models use and produce the exact same text using different numbers of tokens, which means a lower price per token does not necessarily mean a lower bill. Imagine two identical pizzas. One is cut into 8 slices at $2 each. The other is cut into 16 slices at $1.25 each. The second place advertises cheaper slices, but the whole pizza costs $20 instead of $16. Bummer ... your stomach doesn't actually care about the number of slices you just ate. I know you are hungry now, but back to tokens. In one small comparison spanning English, technical, multilingual, and numerical text, the tokenizer we use for GPT-5.6 Sol used 766 tokens versus an estimated 1,170 for Claude Opus 5. That's a very significant difference of about 34.5% fewer tokens. You can get the same exact text, but pay for all those extra tokens. The price per token doesn't really tell this story. Even correcting for tokenizer differences misses the bigger point. What actually matters is price per successful outcome, and for that you can use benchmarks as a starting point, but really you have to try it and measure on your own use cases. That's all. May the tokens flow.

返信 832 · リポスト 337 · いいね 6,892

返信先 @RJMcGirr

@RJMcGirr Sol can probably migrate you in a few hours. Why isn't it worth the many $$$ you would get back in return?

返信 15 · リポスト 0 · いいね 198

返信先 @gusfeliciano

@gusfeliciano We have ZDR, tell me more

返信 10 · リポスト 1 · いいね 197

返信先 @thsottiaux

@ns123abc I'll write more extensively about this, but OpenAI token != A\ token. We use different tokenizers. And OpenAI's tokenizer is significantly more efficient. This is important given you pay per token over the API and in usage. The difference is ~30% in number of tokens, definitely https://t.co/RwDz5AdK1t

返信 26 · リポスト 10 · いいね 252

返信先 @SnoyIatK

@SnoyIatK Which partnership programs do we need?

返信 28 · リポスト 2 · いいね 156

返信先 @_hibana01

@_hibana01 Yes

返信 3 · リポスト 0 · いいね 159

返信先 @ns123abc

@ns123abc Talking about sticker price?

返信 13 · リポスト 1 · いいね 553

返信先 @drgrandehead

@drgrandehead @nxc42 DM

返信 5 · リポスト 0 · いいね 67

返信先 @DustdrifterX

@DustdrifterX DM

返信 3 · リポスト 0 · いいね 156

返信先 @AbdoKerdawy

@AbdoKerdawy Sol supports 1M, tell me more?

返信 145 · リポスト 5 · いいね 790

返信先 @AISupplyGuy1

@AISupplyGuy1 Do you have evals? Quite surprising as I would expect for most things you would net out 40-50% cheaper per task with Sol. Can you share more?

返信 41 · リポスト 1 · いいね 404

返信先 @nxc42

@nxc42 DM

返信 5 · リポスト 0 · いいね 218

If you have a company and still use Opus instead of Sol, why so? Does price not matter to you? What's something that would convince you to switch?

返信 4,212 · リポスト 179 · いいね 9,064