There is one increasingly popular way to improve a reasoning model these days: let the model teach itself.
The recipe is called on-policy self-distillation, or OPSD. A student model samples a reasoning trajectory. The teacher is a frozen copy of that same student, except the teacher holds one extra card: the worked solution to the problem. The student never sees that card. The teacher scores the student's response prefix by prefix, and the student updates against those per-token targets.
The logic sounds airtight: the teacher saw the answer, so the student should learn more.
A paper from Meta and UCLA tore that logic apart and reached an uncomfortable conclusion: the teacher peeking at the answer did not teach the answer to the student.
The title is the finding itself — Privileged Information Does Not Guarantee Privileged Guidance, arXiv 2611.16064, by Haoyi Qiu (Meta Superintelligence Labs and University of California, Los Angeles, work done during an internship at Meta) and Akshat Agarwal (Meta Superintelligence Labs). Version 1 appeared on 11 November 2026, and the paper's own comment field says it is work in progress. Full text: https://arxiv.org/abs/2611.16064
What the recipe actually looks like
The experiments run at three Qwen3 scales: 1.7B, 4B and 8B. Evaluation uses five competition-math benchmarks: AIME 2024, AIME 2025, AIME 2026, HMMT November 2025 and HMMT February 2026. Training data is the mathematical-reasoning subset of OpenThoughts, fixed at 3,200 examples for 100 updates at a global batch size of 32, with the student sampling one response of up to 1,024 tokens per problem. Teacher and student score the same sampled prefixes, the loss is forward-KL with each vocabulary term capped at 0.05, only the student's LoRA parameters are updated, the framework is veRL, and the runs sit on NVIDIA H100 GPUs with three seeds matched across conditions.
The audit splits the pipeline into four stages: Source (whether the teacher can actually use the solution), Encoding (whether the resulting token targets are useful), Transmission (whether the student learns from them) and Expression (whether the student applies the behaviour without the solution in hand).
Stage one leaks: with the answer in hand, the teacher still cannot spot the student's mistakes
Start with Source. The paper screens Hendrycks MATH Level 5 for problems the student gets wrong in all 12 sampling attempts. Out of 2,242 cleaned questions, that yields 111 questions (5.0%) at 1.7B, 44 (2.0%) at 4B and 35 (1.6%) at 8B. GPT-5.6-Sol and Claude Opus 4.8 judge whether each problem and its worked solution are valid.
Given the privileged solution, the teacher correctly flags only between 13.0% and 25.5% of erroneous prefixes — and at 4B and 8B that rate is lower than when the teacher sees the problem alone. Pair accuracy, where both judgements must be right, ranges from 6.5% to 13.0%. For overall prefix accuracy, only the 1.7B gain of +5.75pp has a paired 95% interval excluding zero, and even there accuracy stays below the 50% baseline of accepting every prefix.
In other words: holding the reference solution, the teacher still cannot reliably identify which step the student got wrong.
Where the gains actually come from: thinking mode, not the answer
The next section is the most striking. The paper runs privileged-solution teachers and problem-only teachers side by side. Under Thinking-OFF evaluation at Step 100, students trained with a problem-only teacher score a higher mean Avg@12 at all three scales — 1.36pp higher at 1.7B, 15.21pp at 4B and 8.60pp at 8B, with the 1.7B interval including zero.
Repeat training for 16 epochs and the pattern sharpens: at 4B and 8B, problem-only teachers raise Thinking-OFF Avg@12 by 15.98 to 22.70pp, versus 5.68 to 10.66pp for privileged-solution teachers, and the problem-only side wins four of six scale-and-question-set comparisons. In the one-pass runs the privileged-solution teacher leads by 1.02pp to 1.91pp at Step 20, but that relative advantage reverses by Step 100.
The paper names this pattern reasoning arbitrage. What drove the gains was not the solution. It was the teacher's own thinking mode.
The paper does not overclaim
The restraint here matters: it never concludes that privileged solutions are useless. The opposite evidence sits right next to it. Table 1 shows the privileged teacher taking a large lead on hard problems at Avg@4: 40.54% against 2.93% at 1.7B, 64.20% against 1.14% at 4B, 58.57% against 0.00% at 8B. And Purified OPSD, a method the paper uses as a matched control, genuinely works: at Step 100 it scores 0.74 to 1.86pp above Base across Qwen3 scales, while Standard OPSD sits 1.92 to 3.57pp below Base. On Qwen3-4B coding, Purified exceeds Standard by 4.45pp on LiveCodeBench v6 and 6.57pp on CodeContests.
What the paper does insist on is precision: an accuracy gain does not establish that privileged information transferred. Whether the teacher can use the solution, whether its token targets point toward correctness, whether the student learns, and whether the student uses any of it without the solution present — these four have to be measured separately.
Three other groups working on privileged distillation in mathematics reach compatible conclusions. The OP²SD paper (https://arxiv.org/abs/2608.09228) swaps the reference solution for the problem and solution of a different example; across three models and three mathematics benchmarks it still improves over the base model and stays competitive with OPSD. Shrestha and Tessier (https://arxiv.org/abs/2608.18271) find the correct reference gives no consistent performance benefit across teacher generation modes, model sizes and training datasets, and that another problem's solution can beat the correct one on several mathematical reasoning benchmarks. Zhang and colleagues (https://arxiv.org/abs/2609.20612) build the 5,319-problem AMPLE-Math suite with six reasoning views and matched reference-free distillation controls, finding that reference-free distillation already accounts for much of Qwen3-1.7B's improvement under thinking-enabled evaluation, with complete traces adding two percentage points only in SmolLM3-3B at step 50.
Four sets of authors, none of them the same, asking the same question: when you give the teacher a cheat sheet, what exactly does the student learn?
So what
For anyone training models, the practical lesson is blunt: what looks like optimising more information may actually be optimising the model teaching its own strengths back to itself.
The bigger effect is on how results get attributed. When a method gains 5 points, the instinct is to credit the new dataset, the new prompt, the new solution. The design in this paper is worth copying: build a control that differs only in whether the teacher gets the cheat sheet, hold everything else fixed, then test each stage of the pipeline separately.
That is also why the title leans on such a hard word — Guarantee. The finding is that there is no guarantee. Having the information and actually transmitting it are two different things, and between them sits a gap you have to measure.
For practitioners the test is simple. If you are training with a stronger teacher and a harder signal, take the teacher's cheat sheet away, rerun, and see what is left. Whatever survives is what you were actually training.