Fronset Benchmark

Thinking Machines Inkling Small at a glance

Good enough on 27/54 tasks at the 90% bar. Best value on 3 tasks. Best fit for: Long-form Content Generation, Relevance, Classification & Matching, Social & Promotional Content.

Provider
OpenRouter
Model name
thinkingmachines/inkling-small
Cost mode:
Qualifies on
27 / 54 tasks (at 90% bar)
Best value on
3 tasks

Cost vs quality across all tasks

0%25%50%75%100%678910Quality score (6–10)Cost-efficiency vs best value (1.0 = best value)Public Response Generation — quality 8.07, 31.4x the cost of the best-value good-enough optionVoice Profile Generation — quality 8.74, 10.6x the cost of the best-value good-enough optionModel Specific Prompt Adaptation — quality 8.21, 14.5x the cost of the best-value good-enough optionPromotional Message Generation — quality 8.18, 26.3x the cost of the best-value good-enough optionReport Outline Generation — quality 9.46, this model IS the best-value good-enough option★Atomic Fact Claim Extraction — quality 7.32, 18.6x the cost of the best-value good-enough optionGenerated Content Relevance Scoring — quality 8.23, 9.0x the cost of the best-value good-enough optionBatch Text Translation — quality 8.48, 27.7x the cost of the best-value good-enough optionBench Hypothesis Post Generation — quality 8.25, 20.6x the cost of the best-value good-enough optionContent To Image Prompt Generation — quality 8.68, 12.4x the cost of the best-value good-enough optionFactual Claim Refinement — quality 8.31, 13.2x the cost of the best-value good-enough optionStructured Content Summarization — quality 8.32, 10.9x the cost of the best-value good-enough optionResearch Query Generation — quality 7.36, 18.6x the cost of the best-value good-enough optionLanguage Identification — quality 9.84, 8.4x the cost of the best-value good-enough optionBench Article Hypothesis Generation — quality 8.38, 1.0x the cost of the best-value good-enough optionProfile Pool Matching — quality 7.85, 13.1x the cost of the best-value good-enough optionEngagement Opportunity Triage — quality 8.79, 24.1x the cost of the best-value good-enough optionContent Set Relevance Scoring — quality 6.64, 5.2x the cost of the best-value good-enough optionTopic Section Assignment — quality 8.03, 8.0x the cost of the best-value good-enough optionSocial Post Portfolio Selection — quality 8.70, 24.7x the cost of the best-value good-enough optionMulti Perspective Decision Synthesis — quality 8.46, 10.2x the cost of the best-value good-enough optionContent Domain Suggestion — quality 7.59, this model IS the best-value good-enough option★Topic Cluster Labeling — quality 8.29, 8.8x the cost of the best-value good-enough optionResearch Region Identification — quality 8.50, this model IS the best-value good-enough option★Publication Title Package Generation — quality 8.49, 18.5x the cost of the best-value good-enough optionVisual Theme Configuration Generation — quality 8.87, 14.9x the cost of the best-value good-enough optionCatalyst And Scenario Analysis — quality 7.95, 8.2x 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 90% bar are plotted.

Per-task breakdown

TaskCategoryQuality (% of best)ConfidenceOverpay
Report Outline Generation ★Long-form Content Generation98%RANKEDbest value
Research Region Identification ★Structured Data & Fact Extraction96%HIGHbest value
Content Domain Suggestion ★ bestRelevance, Classification & Matching100%MEDIUMbest value
Bench Article Hypothesis GenerationEvidence-Grounded Hypothesis Generation96%MEDIUM1x
Content Set Relevance ScoringRelevance, Classification & Matching95%MEDIUM5.2x
Topic Section AssignmentTopic Organization & Clustering98%MEDIUM8x
Catalyst And Scenario AnalysisFinancial Analysis & Trading Decisions91%MEDIUM8.2x
Language IdentificationRelevance, Classification & Matching97%RANKED8.4x
Topic Cluster LabelingTopic Organization & Clustering93%MEDIUM8.8x
Generated Content Relevance ScoringRelevance, Classification & Matching92%MEDIUM9x
Multi Perspective Decision SynthesisFinancial Analysis & Trading Decisions96%HIGH10x
Voice Profile GenerationLong-form Content Generation91%MEDIUM11x
Structured Content SummarizationContent Summarization & Synthesis93%MEDIUM11x
Content To Image Prompt GenerationInfrastructure & Utility98%MEDIUM12x
Profile Pool MatchingRelevance, Classification & Matching90%MEDIUM13x
Factual Claim RefinementInfrastructure & Utility96%HIGH13x
Model Specific Prompt AdaptationInfrastructure & Utility92%MEDIUM14x
Visual Theme Configuration GenerationLong-form Content Generation99%HIGH15x
Publication Title Package GenerationContent Summarization & Synthesis93%RANKED18x
Atomic Fact Claim ExtractionStructured Data & Fact Extraction90%MEDIUM19x
Research Query GenerationInfrastructure & Utility94%MEDIUM19x
Bench Hypothesis Post GenerationSocial & Promotional Content93%MEDIUM21x
Engagement Opportunity TriageRelevance, Classification & Matching99%RANKED24x
Social Post Portfolio SelectionRelevance, Classification & Matching100%HIGH25x
Promotional Message GenerationSocial & Promotional Content95%MEDIUM26x
Batch Text TranslationInfrastructure & Utility92%MEDIUM28x
Public Response GenerationSocial & Promotional Content95%MEDIUM31x

Overpay — how much more you pay by running this model instead of the best-value model that clears the quality bar 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.