Tystros @Tystros

Joined July 2014

Tweets

40
Followers

11
Following

247
Likes

2K

Tystros @Tystros

2 weeks ago

@thsottiaux My only issues are that I keep repeatedly getting "Selected model is at capacity" during Sol Ultra runs, and that the Codex CLI still feels much less good to use than Claude Code. Like the subagent management during Fable Ultracode is much better UX in Claude Code.

0 0 0 125 0

View Details

Tystros @Tystros

2 weeks ago

@thsottiaux "Selected model is at capacity. Please try a different model." just happened in the middle of a difficult prompt on Ultra

0 0 0 313 0

View Details

Tystros @Tystros

3 weeks ago

@GregKamradt @arcprize but do you still apply the 10K cap in the future if you know the pricing in advance? so if you had known the price of 5.6 before, would it basically be stuck at a lower score because it couldn't have gone above 10K in price?

0 0 4 913 0

View Details

Tystros @Tystros

3 weeks ago

@GregKamradt I really want to know ARC AGI 3 results for Fable vs GPT 5.6

0 1 2 70 0

View Details

Tystros @Tystros

2 months ago

@GregKamradt @leothecurious I don't get why you can't just require the frontier labs to give you some free API usage for a benchmark run, and remove any cost cap. if a model could achieve 20% on ARC AGI 3 but cost 50k USD, I would want to know, and not only see the results capped at 10k.

0 0 0 70 0

View Details

Tystros @Tystros

2 months ago

@GregKamradt I still don't understand how you will deal with the price, with the 10k USD cap you have on ARC AGI 3. If Opus 4.8 already hits the 10k USD cap, how can you ever benchmark Fable?

1 0 6 2K 1

View Details

Tystros @Tystros

2 months ago

@GregKamradt @arcprize You really need to fix the X axis with the costs. When the Mythos-class model comes out that will probably be 5x Opus cost, how is that supposed to ever show up correctly on the leaderboard? It seems in the near future models will only become more and more expensive.

0 0 0 88 0

View Details

Tystros @Tystros

3 months ago

@thsottiaux When a session reaches a certain size, I always get the "Falling back from WebSockets to HTTPS transport. stream disconnected before completion: websocket closed by server before response.completed" error and that's super annoying, as then it continues much slower.

0 0 1 326 0

View Details

Tystros @Tystros

3 months ago

@4m473r45u @GregKamradt it's not just one game, its 135 unique games. so at 10K total cost its $74 per game, if it would include all games.

0 0 4 266 0

View Details

Tystros @Tystros

3 months ago

@GregKamradt what does that mean in practice? that the models did not attempt all games, because you stopped as soon as the cost reached 10K, which might be after only half the games were attempted?

1 0 0 330 0

View Details

Tystros @Tystros

3 months ago

@GregKamradt thats a big downgrade over ARC AGI 1/2 though. I'd love to know how the intelligence of the models scales with reasoning levels. Also, why do both models show up as costing exactly the same (10K) on the leaderboard? Is that a leaderboard UI bug?

1 0 0 378 0

View Details

Tystros @Tystros

a year ago

@SaveOnSend @kirtan what comparison website is that?

1 0 0 15 0

View Details

Tystros @Tystros

2 years ago

@SaveOnSend @HSBC @Wise @Atlantic_Money @remitly Wise isn't "cheap", Wise has a 0.5% fee which quickly adds up to a lot for larger amounts. Remitly is even more expensive with a ~1% fee.

1 0 0 62 0

View Details

Tystros @Tystros

2 years ago

@SaveOnSend @HSBC @Wise @Atlantic_Money @remitly It's just really unfortunate that there isn't a single service yet that offers cheap USD->EUR exchanges for people in the whole EU or US, or is there?

1 0 0 72 0

View Details

Tystros @Tystros

2 years ago

@RaMansell @The_DarkPill @MWSieb @memcculloch if your drive really works and could revolutionize space travel from one second to the other, don't you think you should open source all "proprietary information" immediately for the sake of humanity, instead of keeping it private just for the benefit of a few?

1 0 0 25 0

View Details

Tystros @Tystros

2 years ago

@RaMansell @_InsanityDemon_ @memcculloch @3DX3EM @ivo_ltd why can you not just make everything you have public and allow anyone to independently verify the results you're seeing?