Thinking Machines LabInkling
Inkling
Thinking Machines Lab · Inkling. Released Jul 15, 2026.
Inkling is a model from Thinking Machines Lab in the Inkling family, released Jul 15, 2026. evals.report tracks 14 reported Inkling benchmark scores across Humanity's Last Exam, AIME 2026, GPQA Diamond, SWE-bench Verified, SWE-bench Pro, Terminal-Bench 2.1, MCP Atlas, BrowseComp, and 6 more — each shown with its benchmark, metric, source status, and date, and never combined into a single ranking.
Open14 results
Benchmark results 14
Compare this model| Benchmark | Category | Score | Metric | Status | Date | |
|---|---|---|---|---|---|---|
| Humanity's Last Exam | Reasoning | 29.7% | accuracy | Verified | Jul 15, 2026 | Details |
| AIME 2026 | Reasoning | 97.1% | accuracy | Verified | Jul 15, 2026 | Details |
| GPQA Diamond | Reasoning | 87.2% | accuracy | Verified | Jul 15, 2026 | Details |
| SWE-bench Verified | Coding | 77.6% | % resolved | Verified | Jul 15, 2026 | Details |
| SWE-bench Pro | Coding | 54.3% | % resolved | Verified | Jul 15, 2026 | Details |
| Terminal-Bench 2.1 | Agents | 63.8% | task success | Verified | Jul 15, 2026 | Details |
| MCP Atlas | Tool use | 74.1% | pass rate | Verified | Jul 15, 2026 | Details |
| BrowseComp | Agents | 77.1% | accuracy | Verified | Jul 15, 2026 | Details |
| SimpleQA Verified | Other | 43.9% | accuracy | Verified | Jul 15, 2026 | Details |
| AA-Omniscience: Knowledge and Hallucination Benchmark | Reasoning | 2.1 | AA-Omniscience Index | Verified | Jul 15, 2026 | Details |
| IFBench | Reasoning | 79.8% | accuracy | Verified | Jul 15, 2026 | Details |
| Global-MMLU | Reasoning | 88.7% | accuracy | Verified | Jul 15, 2026 | Details |
| MMMU-Pro | Multimodal | 73.5% | accuracy | Verified | Jul 15, 2026 | Details |
| CharXiv | Multimodal | 78.1% | accuracy | Verified | Jul 15, 2026 | Details |