From f06860494848db080c9a80a0ffa83203b042056b Mon Sep 17 00:00:00 2001 From: Andrej Karpathy Date: Tue, 10 Mar 2026 06:26:39 +0000 Subject: [PATCH] new leaderboard entry coming from improvements of autoresearch round 1, time to gpt-2 from 2.02 hours to 1.80 hours --- README.md | 1 + dev/LEADERBOARD.md | 5 +++++ 2 files changed, 6 insertions(+) diff --git a/README.md b/README.md index 077fd9c..1fed675 100644 --- a/README.md +++ b/README.md @@ -18,6 +18,7 @@ Presently, the main focus of development is on tuning the pretraining stage, whi | 2 | 2.91 | 0.74504 | 0.2578 | d26 slightly undertrained **+fp8** | Feb 2 2026 | a67eba3 | @karpathy | | 3 | 2.76 | 0.74645 | 0.2602 | bump total batch size to 1M tokens | Feb 5 2026 | 2c062aa | @karpathy | | 4 | 2.02 | 0.71854 | 0.2571 | change dataset to NVIDIA ClimbMix | Mar 4 2026 | 324e69c | @ddudek @karpathy | +| 5 | 1.80 | 0.71808 | 0.2690 | autoresearch [round 1](https://x.com/karpathy/status/2031135152349524125) | Mar 9 2026 | 6ed7d1d | @karpathy | The primary metric we care about is "time to GPT-2" - the wall clock time needed to outperform the GPT-2 (1.6B) CORE metric on an 8XH100 GPU node. The GPT-2 CORE score is 0.256525. In 2019, the training of GPT-2 cost approximately $43,000 so it is incredible that due to many advances over 7 years across the stack, we can now do so much faster and for well below $100 (e.g. at the current ~$3/GPU/hr, an 8XH100 node is ~$24/hr, so 2 hours is ~$48). diff --git a/dev/LEADERBOARD.md b/dev/LEADERBOARD.md index 556ec3c..f20d455 100644 --- a/dev/LEADERBOARD.md +++ b/dev/LEADERBOARD.md @@ -191,3 +191,8 @@ Mean is 0.25714 (higher than the GPT-2 threshold needed), max-min is 0.01646. So NOTE: The `val_bpb` is as of this run *NOT* comparable due to the data distribution change to the previous 3 runs. This run happens to be at `0.71854` validation bpb. If the dataset is not changed, the `val_bpb` number is a great, smooth metric to track relative performance w.r.t. and has less noise than CORE. +## Run 5 + +Achieved Mar 9, 2026 on commit `6ed7d1d`. Exactly the same launch command as Run 4 except `--target-param-data-ratio=8.7`. I ran 5 identical runs, the average CORE was 0.2690, which is quite a bit above the needed threshold of 0.2565. But the reason I didn't decrease the ratio further (i.e. train shorter) is that while the CORE "safety gap" is large, the val_loss safety gap is smaller - 0.71808, which we want to be below the Run 4 val loss of 0.71854. It's likely that we could have reduced the ratio even lower, possibly to 8.6, but it's not worth splitting hairs at this point. + +This commit is special because all of the improvements that went into [this commit](https://github.com/karpathy/nanochat/commit/6ed7d1d82cee16c2e26f45d559ad3338447a6c1b) came from fully autonomous "research" done by a private version of [autoresearch](https://github.com/karpathy/autoresearch) run on a d12 model. I wrote more about this in [this tweet](https://x.com/karpathy/status/2031135152349524125). The changes easily translated from d12 to d24, hence new leaderboard record, taking us from 2.02 hours "time to GPT-2" to 1.80 hours.