αalphaBell / observatoryExplore the Observatory

Release lens 02 / source-pinned

Same published blueprint.
New reported scores.

DeepSeek V4.1 Flash 1017 and its predecessor publish byte-identical model and inference configurations. The 1017 model card reports higher scores on every one of its 11 paired benchmarks. Explore each pair below.

Reading the evidence Scores come from DeepSeek's model card; alphaBell has not rerun the benchmarks. Equal configuration files do not establish equal weights or explain the changes.

11/11

paired scores higher in 1017

+11.1

largest reported score-point gain

2/2

published configuration files identical

01 / Reported benchmarks

What changed on the card?

Compare the two releases within each row. Each benchmark has its own evaluation method, so scores across rows are not directly comparable.

Original1017Scale: reported score, 0 to 100

HLE's text-only subset is reported separately in the source: 39.1 for the original and 44.0 for 1017. Terminal-Bench 4.0 uses a Claude Code harness; DeepSWE v1.1 uses Mini-SWE-V2. All figures are publisher-reported.

02 / Configuration check

What stayed in the published files?

The pinned releases have the same byte-level SHA-256 digest for each of these two files. Expand a row to inspect both exact source links and the full digest.

Model configIdentical bytes
8be45ce0476004a3f529fd896115a4a2e800a129ad2d3ec05b16050f52e21879
Inference configIdentical bytes
2e84f45cf1dac8c7fcbb200e96667d4b913275690668ed496f24c7747207a809

A matching config describes the same published structure and settings. It says nothing about whether the weights, training data, post-training process or benchmark setup changed.

Pinned sources

Check the primary evidence