AI systems are improving at choosing next steps in AI research.
This assessment rests in part on evidence not shown: at least one of its sources reports a result from data it has not published.
Assessment
The evidence favors the claim, but the examination is incomplete or the evidence is indirect.
The claim holds that AI systems are getting better at deciding what to do next in AI research: which experiment to run, which lead to follow, when to change course. That is a narrower and harder skill than writing code or running a fixed experiment, and it is the one most often named as the remaining human advantage.
Two kinds of evidence, from different hands and by different methods, point the same way. Anthropic reported in 2026 that on 129 moments taken from its researchers' real debugging and investigation sessions, the share of cases where a model's proposed next step was judged better than the researcher's rose from 51% for Claude Opus 4.5 (November 2025) to 64% for Claude Mythos Preview (April 2026). The test used unpublished data, was judged by another Claude model that knew how each session ended, and deliberately picked moments where the human choice was poor, so it shows improvement between models rather than parity with human judgment, and it comes from a developer with a stake in its models looking capable. The external TasteVal benchmark (Jaffe and Sherburn, October 2026), in which the evaluated model chooses experiments while a fixed coding agent runs them, found the efficiency of frontier models relative to human experts rising across 20 models from 2023 to 2026, with the trend steepening since late 2025 and the best model exceeding the expert baseline. Several other 2025-2026 studies of forecasting which research ideas will work report similar gains.
The evidence favors the claim, but it is indirect: each test captures a narrow slice of research judgment, the tasks are mostly unreleased, and model-graded judging can drift. Neither test reaches the broadest form of the skill, choosing which problems are worth working on at all, where even the developers reporting gains say humans still lead.
Full reasoning: the evidence and decisions behind this verdict
Prior: given rapid general capability gains in 2024-2026 and that next-step choice is a component of agentic task performance (which has improved sharply), improvement on the narrow skill is very likely; the open question is mainly whether measurements isolate it.
Evidence read: (1) Anthropic Institute, "When AI builds itself" (read in full around the relevant section). Internal next-step test, n=129 selected detours, Jan-Mar 2026 sessions, judged by a separate Claude with hindsight; 51% to 64% between Opus 4.5 and Mythos Preview; control set of strong human moves gave about 20%. Evidence not shown; developer interest recorded. Two data points only. (2) TasteVal, arXiv 2610.06824 (abstract read): 8 unreleased AI R&D tasks, model as Researcher with fixed Coder, 20 models 2023-2026, compute multiplier doubling about every 3 months since Dec 2025 (95% CI 1.7-5.0), Opus 5.5 at 2.3x the expert baseline (CI 1.15-4.37); final normalized performance doubles every 14.6 months. Independent of the Anthropic test in method, though tasks are unreleased and the authors' affiliation was not checked. (3) Search also surfaced several idea-outcome forecasting papers (e.g. arXiv 2506.00794, 2605.21491) reporting LMs beating or approaching experts at predicting which research idea performs better; not read in full, weighted lightly.
What would change this: evidence that the gains vanish when coding ability is controlled for, that model judges favor newer-model outputs, or replication failures of TasteVal-style trends. Status supported rather than verified because the measures are narrow and largely unverifiable externally.
Decomposition
This claim is atomic: it bottoms out in a bedrock fact, an open empirical question, or a value premise, and does not decompose further.
Provenance
Where this claim has been said, linked to its canonical form.
Neither source has been traced to an origin, or to one another.
On TasteVal, the compute multiplier of frontier models has doubled approximately every 3.0 months since December 2025 (95% CI 1.7-5.0), up from every 14 months between 2023 and December 2025.
Benchmark of 8 unreleased AI R&D tasks where the evaluated model acts as Researcher choosing experiments while a fixed Coder implements them; 20 models from 2023-2026 compared with 24 human experts.
We view this result as an early signal that AI systems are getting better at making the kinds of judgement calls that AI research depends on.
Next-step selection test on 129 real research-session detours, judged by a separate Claude.
The source reports a result from evidence it holds and has not published. The report describes an internal test on 129 detours from researchers' Claude Code sessions, January to March 2026, scored by a separate Claude that could see how each session ended. The win rate against the researcher's choice rose from 51% to 64% between two models. The sessions, the judging prompts and per-case results are not published, and no outside party is reported to have examined them. The moments were selected because the human choice was poor, so the figures compare models with each other, not with human judgment in general; the report says so itself.
Cite this claim: a formal citation with its evidence attached
Embed this claim: show it on another site with its current assessment
An embed reports the graph's judgment as it stands when the page is viewed: the status with its confidence, how disputed the claim is, and the date. It links back here.
<blockquote class="minerval-claim" data-claim-id="f5a654f9-9ef3-4144-9869-6bff1c6d9631"><a href="https://minerval.ai/claims/f5a654f9-9ef3-4144-9869-6bff1c6d9631">AI systems are improving at choosing next steps in AI research.</a></blockquote> <script async src="https://minerval.ai/embed.js"></script>
<iframe src="https://minerval.ai/embed/claims/f5a654f9-9ef3-4144-9869-6bff1c6d9631" title="Minerval: AI systems are improving at choosing next steps in AI research." width="480" height="200" style="border:0;max-width:100%" loading="lazy"></iframe>
[](https://minerval.ai/claims/f5a654f9-9ef3-4144-9869-6bff1c6d9631)
The badge now: 
Contribute
Every judgment on this page is open to challenge. A contribution is evaluated on its merits by the reviewer; if it succeeds the page changes, and if it does not, the reasons are stated. Either way the exchange becomes part of the claim’s public record.
Stated views None yet. A view stated to the graph is shown on this page in its speaker’s words, as a source like any other, and carries no weight in the assessment.
The attention this claim received was paid for by a funded mandate. Funding buys only scheduling: it can make an assessment happen sooner, or reach deeper into a subtree. It has no influence on what the assessment concludes, and none on which claims enter the graph; assessments run under the same public standards whoever pays, funders never see or shape a verdict before anyone else, and mandates that attempt to steer conclusions are refused.
Created by extractor · Nov 13, 2026. Every judgment on this page is accompanied by a reasoning trace.