BlogLLM Benchmark Scores as Published, Side by Side
Inkling: Benchmark Scores as Published
Every result the catalog holds for Inkling: 49 tests in 13 areas, from 4 pages. 24 printed by Thinking Machines itself, the rest by the company named beside each number. Each number is as printed; nothing is ranked or rescored.
Developer Thinking MachinesReleased 2026-07-14Input Text, image, audioOutput TextWeights Open (Apache 2.0)Data version 2026103001, as of 2026-10-30
Knowledge Work
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| AutomationBench | 12.8% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: xhigh |
| GDPval-AA v2 | 1233 | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15; printed on 1 more page | Effort: 0.99 |
| APEX-Agents | 19% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: xhigh |
| JobBench | 31.1 | Meta, Meta Muse Spark Open 816B-A42B model cardchecked 2026-10-19 | Effort: xhigh |
| GDPval-AA (Elo) | 1234 | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: xhigh |
Agentic Coding
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| SWE-bench Verified | 77.6% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15 | Effort: 0.99 |
| SWE-Bench Pro | 54.3% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15; printed on 1 more page | Effort: 0.99 |
| Terminal-Bench 2.1 | 63.8% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15; printed on 2 more pages | Effort: 0.99; Harness: best harness per model, as printed |
| SWE-Atlas-QnA (strict) | 25.5% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: xhigh; Tools: no internet (strict) |
Science and Math
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPQA Diamond | 87.2% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15; printed on 2 more pages | Effort: 0.99 |
| SciCode | 46.1% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15; printed on 1 more page | Effort: xhigh |
| AIME 2026 | 97.1% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15 | Effort: 0.99 |
| CritPt | 5.4% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: xhigh |
Long Context
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| AA-LCR | 73.3% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: xhigh |
Computer Use
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| OSWorld-Verified | 40.1% | Meta, Meta Muse Spark Open 816B-A42B model cardchecked 2026-10-19 | Effort: xhigh |
Multimodal Understanding
General Knowledge and Reasoning
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Humanity's Last Exam (no tools) | 29.7% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15; printed on 1 more page | Effort: 0.99; Tools: none |
| Humanity's Last Exam (with tools) | 46% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15; printed on 1 more page | Effort: 0.99; Tools: with tools |
| AA-Omniscience Accuracy | 42% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: xhigh |
| SimpleQA Verified | 43.9% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15 | Effort: 0.99 |
| AA-Omniscience Index | 1 | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15; printed on 1 more page | Effort: 0.99 |
| AA-Omniscience Non-Hallucination | 32% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: xhigh |
| Humanity's Last Exam Diamond (no tools) | 11.2% | Meta, Meta Muse Spark Open 816B-A42B model cardchecked 2026-10-19 | Effort: xhigh; Tools: none |
Tool Use and Workflows
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| τ³-bench Banking | 23.7% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15; printed on 1 more page | Effort: 0.99 |
| Toolathlon Verified | 45.5% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15; printed on 1 more page | Effort: xhigh |
| MCP Atlas | 74.1% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15; printed on 1 more page | Effort: 0.99 |
| MCPMark | 51.2% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: xhigh |
| WildClawBench | 52.3% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: xhigh |
Web Search and Research
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| BrowseComp | 77.1% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15; printed on 1 more page | Effort: 0.99; Harness: with context (w/ Ctx), as printed |
Instruction Following
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| IFBench | 79.8 | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15 | Effort: 0.99 |
Multilingual
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Global-MMLU-Lite | 88.7% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15 | Effort: 0.99 |
Safety and Alignment
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| FORTRESS (Adversarial) | 78% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15 | Effort: 0.99 |
| FORTRESS (Benign) | 95.9% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15 | Effort: 0.99 |
| StrongREJECT | 98.6% | Thinking Machines, Thinking Machines Inkling model cardchecked 2026-10-15 | Effort: 0.99 |
| FORTRESS adversarial (static), ARSlower is better | 25 | Meta, Meta Muse Spark Open 816B-A42B model cardchecked 2026-10-19 | Effort: xhigh |
| FORTRESS adversarial (static), ORSlower is better | 7.4 | Meta, Meta Muse Spark Open 816B-A42B model cardchecked 2026-10-19 | Effort: xhigh |
| StrongREJECT v2 (adaptive)lower is better | 86.4 | Meta, Meta Muse Spark Open 816B-A42B model cardchecked 2026-10-19 | Effort: xhigh |
Preference and Composite Rankings
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Artificial Analysis Intelligence Index | 26 | Meta, Meta Muse Spark Open 816B-A42B model cardchecked 2026-10-19 | Effort: xhigh |
Price
API list prices in US dollars per one million tokens, as printed on the developer's pricing page on the day checked. Cost per task is measured by Vals AI.
| Price | Value | Printed by | Setup as printed |
|---|---|---|---|
| Cost per taskVals AI, USD, lower is better | $1.61 | Vals AI, Vibe Code Bench v1.1 leaderboardchecked 2026-10-27 | Harness: OpenHands |
More from Thinking Machines
Every number here is read from the same catalog file as the benchmark table, data version 2026103001. Where one page printed the model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.