Fronset Benchmark

Thinking Machines Inkling at a glance

Good enough on 27/54 tasks at the Balanced step. Best value on 1 task. Best fit for: Financial Analysis & Trading Decisions, Infrastructure & Utility, Social & Promotional Content.

Provider
OpenRouter
Model name
thinkingmachines/inkling
What the steps mean
Cost mode:
Qualifies on
27 / 54 tasks (at the Balanced step)
Best value on
1 tasks

Cost vs quality across all tasks

0%25%50%75%100%78910Quality score (7–10)Cost-efficiency vs best value (1.0 = best value)Factual Claim Refinement — quality 8.05, 25.6x the cost of the best-value good-enough optionResearch Query Generation — quality 7.27, 82.3x the cost of the best-value good-enough optionMulti Perspective Decision Synthesis — quality 8.56, 12.4x the cost of the best-value good-enough optionBench Hypothesis Post Generation — quality 8.29, 41.5x the cost of the best-value good-enough optionEngagement Opportunity Triage — quality 8.87, 101.2x the cost of the best-value good-enough optionPublic Response Generation — quality 7.86, 114.4x the cost of the best-value good-enough optionTopic Cluster Labeling — quality 8.29, 141.0x the cost of the best-value good-enough optionSocial Post Portfolio Selection — quality 8.52, 119.0x the cost of the best-value good-enough optionCatalyst And Scenario Analysis — quality 7.94, 19.5x the cost of the best-value good-enough optionResearch Region Identification — quality 8.37, 2.8x the cost of the best-value good-enough optionProfile Pool Matching — quality 8.07, 44.6x the cost of the best-value good-enough optionContent To Image Prompt Generation — quality 8.77, 40.4x the cost of the best-value good-enough optionModel Specific Prompt Adaptation — quality 8.19, 25.5x the cost of the best-value good-enough optionVoice Profile Generation — quality 8.94, 19.2x the cost of the best-value good-enough optionThematic Topic Discovery — quality 7.82, 2.2x the cost of the best-value good-enough optionStructured Content Summarization — quality 8.29, 55.7x the cost of the best-value good-enough optionNewsletter Copy Generation — quality 8.79, 13.7x the cost of the best-value good-enough optionPersona Policy Eligibility Check — quality 8.48, 37.7x the cost of the best-value good-enough optionPublication Title Package Generation — quality 8.48, 77.8x the cost of the best-value good-enough optionProfiled Document Analysis — quality 8.35, 22.2x the cost of the best-value good-enough optionLanguage Identification — quality 9.68, 74.1x the cost of the best-value good-enough optionPublication Section Taxonomy Generation — quality 8.80, this model IS the best-value good-enough option★Subject Configuration Generation — quality 8.50, 18.9x the cost of the best-value good-enough optionPromotional Message Generation — quality 8.42, 59.2x the cost of the best-value good-enough optionBatch Text Translation — quality 8.59, 45.0x the cost of the best-value good-enough optionBench Article Hypothesis Generation — quality 8.20, 4.7x the cost of the best-value good-enough optionSection Prompt Generation — quality 8.25, 25.7x the cost of the best-value good-enough option

within ~1.3× of the best-value model · 1.3–2× · >2× · ★ this model is the best-value pick on that task. Top-right = best quadrant. Only tasks where this model qualifies at the Balanced step are plotted.

Per-task breakdown

TaskCategoryQuality (% of best)ConfidenceOverpay
Publication Section Taxonomy Generation ★Long-form Content Generation96%RANKEDbest value
Thematic Topic DiscoveryTopic Organization & Clustering95%MEDIUM2.2x
Research Region IdentificationStructured Data & Fact Extraction95%HIGH2.8x
Bench Article Hypothesis GenerationEvidence-Grounded Hypothesis Generation94%HIGH4.7x
Multi Perspective Decision SynthesisFinancial Analysis & Trading Decisions97%HIGH12x
Newsletter Copy GenerationLong-form Content Generation100%RANKED14x
Subject Configuration GenerationFinancial Analysis & Trading Decisions99%HIGH19x
Voice Profile GenerationLong-form Content Generation93%HIGH19x
Catalyst And Scenario AnalysisFinancial Analysis & Trading Decisions90%MEDIUM20x
Profiled Document AnalysisFinancial Analysis & Trading Decisions91%MEDIUM22x
Model Specific Prompt AdaptationInfrastructure & Utility92%MEDIUM26x
Factual Claim RefinementInfrastructure & Utility93%MEDIUM26x
Section Prompt GenerationInfrastructure & Utility91%MEDIUM26x
Persona Policy Eligibility CheckRelevance, Classification & Matching92%MEDIUM38x
Content To Image Prompt GenerationInfrastructure & Utility98%HIGH40x
Bench Hypothesis Post GenerationSocial & Promotional Content94%HIGH42x
Profile Pool MatchingRelevance, Classification & Matching93%MEDIUM45x
Batch Text TranslationInfrastructure & Utility93%MEDIUM45x
Structured Content SummarizationContent Summarization & Synthesis93%HIGH56x
Promotional Message GenerationSocial & Promotional Content97%RANKED59x
Language IdentificationRelevance, Classification & Matching97%MEDIUM74x
Publication Title Package GenerationContent Summarization & Synthesis93%RANKED78x
Research Query GenerationInfrastructure & Utility92%MEDIUM82x
Engagement Opportunity Triage bestRelevance, Classification & Matching100%RANKED101x
Public Response GenerationSocial & Promotional Content93%MEDIUM114x
Social Post Portfolio SelectionRelevance, Classification & Matching98%HIGH119x
Topic Cluster LabelingTopic Organization & Clustering94%MEDIUM141x

Overpay — how much more you pay by running this model instead of the best-value model that meets the quality step on that task (marked ★). "16x" means you overpay 16× — the same output for 16× the best-value good-enough option; ★ means this model is that option (no overpayment). Confidence — how sure we are about the quality score (more judgments + more agreement = higher confidence): RANKED many independent judges scored this model's outputs and their agreement is very high (most confident) — HIGH many judges have scored it and they mostly agree (well-pinned) — MEDIUM enough judges have weighed in to publish, but they disagree more than we'd like (treat with a small grain of salt). LOW-confidence cells are hidden everywhere on the site. See the methodology for the exact thresholds.