Three publicly available LLMs achieve 96% average accuracy on text-only physics problems with no significant differences among models¶
Claim · All claims
Evidence · 1 study · causalr1·q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person · recheck by 2029: tool and model accuracy changes quickly
Subclaims¶
Reading the evidence codes
q is the design tier, i the impact, n the sample, and the kind with its r
the rigour judged within that kind: what each code means.
This claim: no standardized effect is recorded in observations/ for it yet, so no benchmark reading is possible.
q2 i? All three models demonstrated strong performance on text-only physics problems with an average accuracy of 96% (95% CI [0.93, 0.98]). → Akshay Syal 2026
Evidence¶
Akshay Syal 2026¶
Akshay Syal, Lawrence Swaminathan Xavier Prince, Evin Gultepe, Nik Bear Brown, Srinivas Sridhar. (2026). A Dialogue-Based Framework for Correcting Multimodal Errors in AI-Assisted STEM Education. https://arxiv.org/abs/2605.04131 doi:10.48550/arxiv.2605.04131
q2 · i? · causal · r1
Baseline evaluation of GPT-5, Claude Sonnet-4.5, and Gemini-3 Pro Preview on 100 text-only OpenStax physics questions, answers manually verified against the OpenStax key. The article reports "an average accuracy of 96% (95% CI [0.93, 0.98])". Cochran's Q found no significant model differences (Q = 3.6, df = 2, p = 0.17).
"On text-only physics problems (Table 1), all three models demonstrated strong performance with an average accuracy of 96% (95% CI [0.93, 0.98])."
Discussion¶
Related Claims¶
- LLM accuracy drops from 96% on text-only to 74% on multimodal physics problems, with significant differences emerging among models — related
- A three-step structured multimodal dialogue intervention corrected 82% of baseline errors and raised multimodal accuracy to 95%, with visual processing errors corrected at 100% but context misinterpretation errors resistant — related
- The Multimodal Interference Effect: mapping visual representations to physical concepts degrades LLM performance even on non-visual problem aspects — related
- LLM errors on multimodal physics problems fall into four categories, with visual processing errors most prevalent, and error type distribution does not differ significantly across models — related
Page metadata
| Field | Value |
|---|---|
type |
claim |
id |
llms-96-percent-text-only-physics-accuracy |
title |
Three publicly available LLMs achieve 96% average accuracy on text-only physics problems with no significant differences among models |
description |
Three publicly available LLMs achieve 96% average accuracy on text-only physics problems with no significant differences among models |
status |
review |
evidence_strength |
moderate |
newest_evidence |
2026 |
recheck_by |
2029 |
generated |
by process:wiki-ingest on 2026-10-11 |
trust |
unverified — no one has confirmed this page's content |
sources |
1 — akshay-syal-2026 |
Frontmatter is the page's machine-readable record — see the Schema & Guide.