DeepSWE
No official DeepSWE mini-swe-agent row for this identity on the public board we fetched.
Smaller Tinker sibling of Inkling. Official Apache-2.0 repo. AA 41. vals hero 31.86% ±1.10. Full Inkling on the same board is 42.06% ±1.16. OpenRouter $0.45 / $1.20; AA prints $0.30 / $1.20. No official DeepSWE row. Not hosted.
Open Apache-2.0 at thinkingmachines/Inkling-Small. Catalog $0.45 / $1.20, cache read $0.10 /M, 524,288 / 262,144, created unix 1785443117 (30 Jul 2026).
Full Inkling is the other card (AA 42, vals 34.10% on that page: do not copy those tiles here).
Use Inkling Small when you want this cheaper official checkpoint. AA 41; vals hero 31.86% ±1.10. We do not invent a DeepSWE percent. Hub API 19 Aug 2026: 365 likes, 108,016 downloads.
No official DeepSWE mini-swe-agent row for this identity on the public board we fetched.
vals.ai Accuracy Rankings hero, 19 Aug 2026: 31.86% ±1.10. Full Inkling on the same board is 42.06% ±1.16. Do not copy it here. The Inkling card’s own vals hero is 34.10% ±1.13 on a different vals URL.
Ranked view: Best coding models: the independent leaderboard puts this row and every other card that carries an independent coding score on one board, with the confidence intervals left visible.
Copied from AA or vals HTML we opened. 0.0% placeholder bars are omitted. Do not average these into the hero tiles, and do not invent CI, $/task, or steps for Artificial Analysis.
Plus is $25/mo with $25 weekly hosted usage. The Mac app stays free with your own keys. Get Plus is not Download for Mac.
Community quants: Community quants are community.
This is the model repo, not Download for Mac. Downloads and likes from the Hugging Face API on 2026-08-19.
Not listed as a Continuum hosted id. OpenRouter thinkingmachines/inkling-small.
We list first-party and OpenRouter, plus Continuum only when the live public allowlist named the id. This is not a 15-host routing table.
OpenRouter catalog 19 Aug 2026: $0.45 / $1.20, cache read $0.10 /M. AA prints $0.30 / $1.20, 67 tok/s, $192.52 Index eval. Cite both. No DeepSWE $/task.
Artificial Analysis Intelligence Index 41, fetched 19 Aug 2026. Same page: 67 tok/s; $192.52 Index eval. AA $0.30 / $1.20 vs OR $0.45 / $1.20.
| Claim | Source | As of |
|---|---|---|
| OR catalog $0.45 / $1.20, cache $0.10, 524,288 / 262,144, created 1785443117 | OpenRouter catalog | 2026-08-19 |
| AA Index 41; $0.30 / $1.20; 67 tok/s; $192.52 Index eval | Artificial Analysis Inkling Small | 2026-08-19 |
| vals Accuracy 31.86% ±1.10 | vals.ai Accuracy Rankings 2026-08-19 | 2026-08-19 |
| HF likes 365; downloads 108,016; usedStorage 531,944,513,947; safetensors.total 265,956,439,090; Apache-2.0; created 2026-07-27T22:33:44Z | Hugging Face API | 2026-08-19 |
| OpenRouter id thinkingmachines/inkling-small, context 524,288 | OpenRouter /api/v1/models | 2026-08-19 |
| HF downloads 108,016 | Hugging Face API thinkingmachines/Inkling-Small | 2026-08-19 |
No other card in this cluster shares a published DeepSWE official row we can put next to this one. We will not compare on vendor-blog numbers.
Continue through Thinking Machines's model family with Inkling.
No. vals Accuracy for Inkling Small is 31.86% ±1.10. 42.06% ±1.16 is a full-Inkling row on the same 19 Aug Accuracy board. The Inkling card chips 34.10% ±1.13 from its own vals model page.
Lab posts pick a harness, an effort, a split, and sometimes a private eval set. DeepSWE publishes the official mini-swe-agent row with cost, tokens, and steps. Scale SWE-bench Pro only counts when the public shared-harness board has a row we can fetch. A higher lab number is usually a different test, not a better one.
No Continuum hosted id is verified for this model, so this page does not invent a curl target. Use the first-party API or the OpenRouter slug thinkingmachines/inkling-small.