Baseten
@baseten
Oct 23
We ran @mirendil's new base model on Baseten's private codebases; that is, code it's never seen. Mirendil-Base-Large had the lowest bits per byte (a tokenizer-independent measure of perplexity) of any model we tested.

We're excited to be the inference partner to serve Mirendil Mirendil team on this launch!

Oct 23, 2026 · 9:35 PM UTC

9
18
224
49,968
RepliesRelevant
Lucas Beyer (bl16)
@giffmana
Oct 23
Replying to @baseten
@baseten @mirendil so baseten does not contain vllm or sglang or TensorRT-LLM code?
4
0
68
5,847
Stewart Slocum
@stewpervised
14h
Replying to @baseten
@baseten @mirendil Impressive!
0
0
0
631
floating point
@yar_vol
21h
Replying to @baseten
@baseten @mirendil But NVidia guy said this is all simple and low taste slop and LLMs he said can only take commodity results and replicate them (without taste he said).
0
0
1
1,515
Molly Bernfeld
@Mollybernn
Aug 5
Introducing Honen

Turn your company context into interactive courses that self-improve over time

We partnered with NVIDIA to bring AI literacy to 250,000 learners

We're ensuring no person or company gets left behind

Try free at honen.com/s/trynow
94
108
1,996
2,241,518
Sahan Paliskara@sahan_paliskara
22h
Replying to @baseten
@baseten @mirendil this is impressive
0
0
0
1,146
John Kovacs
@jj_kovacs
16h
Replying to @baseten
@baseten @mirendil testing on your own private repos is the bit that makes this worth reading. anything public has probably leaked into someone's pretraining by now, so code nobody has seen is about the only clean held-out set left
0
0
0
408
Tianyi Zhang
@mycharmspace
17h
Replying to @baseten
@baseten @mirendil I think you guys using the same coding model and plan for your dev😆
0
0
2
714
Load more