Thinking Machines
Inkling-Small benchmark results
5 results on 5 benchmarks from 2 sources.
Also listed as: inkling-small_xhigh · Inkling Small (xhigh)
All benchmark results
| Benchmarks | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|
| LMArena Text | 1405 | 158 | Overall, style control on; 35359 votes; 95% CI 1400–1409 | Official leaderboard | Oct 22, 2026 |
| LMArena WebDev | 1409 | 92 | Overall; 5328 votes; 95% CI 1400–1419 | Official leaderboard | Oct 22, 2026 |
| LMArena Vision | 1209 | 87 | Overall, style control on; 8074 votes; 95% CI 1201–1217 | Official leaderboard | Oct 19, 2026 |
| GPQA Diamond | 88.5% | 56 | Epoch AI run, accuracy; effort: xhigh | Independent run | Aug 14, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 90.0% | 71 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: xhigh | Independent run | Aug 14, 2026 |
Where Inkling-Small ranks
Primary setting, one best-sourced score per model (official / independent first); neighbours in brackets.
| Benchmark | Score | Rank | Nearby models | #1 |
|---|---|---|---|---|
| GPQA Diamond | 88.5% | #47 / 217 | Grok 4.3 Beta 88.8% · Muse Spark Open 88.6% · Qwen 3.6 Plus 88.4% · Inkling 88.3% | Gemini 4 Argon96.4% |
| AIME (OTIS Mock AIME 2024–2025) | 90.0% | #56 / 200 | Qwen 3.6 Max (Preview) 91.1% · Qwen3.6 27B 91.1% · Muse Spark 88.9% · GPT-5.4 Mini 88.9% | Gemini 4 Argon100.0% |
| LMArena Text | 1405 | #141 / 392 | Mistral Medium 3.1 (2508) 1408 · Grok 4 fast 1405 · Gemini 2.5 Flash (Sep 2025) 1403 · o1 1402 | Gemini 4 Argon1517 |
| LMArena Vision | 1209 | #75 / 154 | Qwen3-VL-235B-A22B-Instruct 1215 · GPT-4.1 1214 · Claude Opus 4 (Thinking 16K, 20250514) 1208 · Claude Sonnet 4 (Thinking 32K, 20250514) 1207 | Claude Fable 51309 |
| LMArena WebDev | 1409 | #84 / 135 | GPT-5.2 1417 · Inkling 1415 · GPT-5.3-Codex (Codex Harness) 1407 · GLM-5V-Turbo 1402 | Claude Opus 5.51807 |
Other Thinking Machines models
Related on this site
FAQ
Where does Inkling-Small rank best?
GPQA Diamond: #47 of 217 (88.5%); #1 there is Gemini 4 Argon (96.4%).
Where do these Inkling-Small scores come from?
From Epoch AI Benchmarking Hub and LMArena leaderboard dataset. Each row links to the original leaderboard, dataset or report, and vendor-reported numbers are labelled.