The Right-Sized LLM Benchmark Cheapest LLM that's good enough.

Thinking Machines Inkling at a glance

Good enough on 27/52 tasks at the 90% bar. Best value on 1 task. Best fit for: Financial Analysis & Trading Decisions, Infrastructure & Utility, Long-form Content Generation, Social & Promotional Cont

Provider
OpenRouter
Model name
thinkingmachines/inkling
Cost mode:
Qualifies on
27 / 52 tasks (at 90% bar)
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)Public Response Generation — quality 7.93, 41.5x the cost of the best-value good-enough optionReference Preserving Analytical Writing — quality 8.60, 12.8x the cost of the best-value good-enough optionShort Promotional Teaser Generation — quality 8.47, 112.1x the cost of the best-value good-enough optionCatalyst And Scenario Analysis — quality 7.99, 2.9x the cost of the best-value good-enough optionMulti Perspective Decision Synthesis — quality 8.63, 13.8x the cost of the best-value good-enough optionPublication Title Package Generation — quality 8.58, 4.8x the cost of the best-value good-enough optionResearch Region Identification — quality 8.39, 2.2x the cost of the best-value good-enough optionEngagement Opportunity Triage — quality 8.87, 37.2x the cost of the best-value good-enough optionProfile Pool Matching — quality 8.20, 6.2x the cost of the best-value good-enough optionLanguage Identification — quality 9.69, 24.4x the cost of the best-value good-enough optionSocial Post Portfolio Selection — quality 8.58, 88.5x the cost of the best-value good-enough optionVoice Profile Generation — quality 9.03, 12.4x the cost of the best-value good-enough optionProfiled Document Analysis — quality 8.45, 1.9x the cost of the best-value good-enough optionFactual Claim Refinement — quality 7.99, 34.6x the cost of the best-value good-enough optionNewsletter Copy Generation — quality 9.05, 8.3x the cost of the best-value good-enough optionBatch Text Translation — quality 8.59, 36.7x the cost of the best-value good-enough optionSubject Configuration Generation — quality 8.69, 11.5x the cost of the best-value good-enough optionContent To Image Prompt Generation — quality 8.88, 14.4x the cost of the best-value good-enough optionProfiled Document Section Analysis — quality 8.90, 26.4x the cost of the best-value good-enough optionPromotional Message Generation — quality 8.46, 20.7x the cost of the best-value good-enough optionCommunity Content Promotion Generation — quality 7.82, 7.9x the cost of the best-value good-enough optionThematic Topic Discovery — quality 7.77, 9.1x the cost of the best-value good-enough optionPublication Section Taxonomy Generation — quality 8.97, this model IS the best-value good-enough option★Section Prompt Generation — quality 8.38, 11.8x the cost of the best-value good-enough optionTopic Cluster Labeling — quality 8.39, 7.2x the cost of the best-value good-enough optionModel Specific Prompt Adaptation — quality 8.54, 14.6x the cost of the best-value good-enough optionPersona Policy Eligibility Check — quality 8.50, 26.5x 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
Publication Section Taxonomy Generation ★Long-form Content Generation96%RANKEDbest value
Profiled Document AnalysisFinancial Analysis & Trading Decisions97%HIGH1.9x
Research Region IdentificationStructured Data & Fact Extraction95%HIGH2.2x
Catalyst And Scenario AnalysisFinancial Analysis & Trading Decisions91%MEDIUM2.9x
Publication Title Package GenerationContent Summarization & Synthesis94%RANKED4.8x
Profile Pool MatchingRelevance, Classification & Matching93%MEDIUM6.2x
Topic Cluster LabelingTopic Organization & Clustering94%MEDIUM7.2x
Community Content Promotion GenerationSocial & Promotional Content90%MEDIUM7.9x
Newsletter Copy GenerationLong-form Content Generation97%RANKED8.3x
Thematic Topic DiscoveryTopic Organization & Clustering91%MEDIUM9.1x
Subject Configuration GenerationFinancial Analysis & Trading Decisions98%HIGH11x
Section Prompt GenerationInfrastructure & Utility98%MEDIUM12x
Voice Profile GenerationLong-form Content Generation92%HIGH12x
Reference Preserving Analytical WritingLong-form Content Generation94%MEDIUM13x
Multi Perspective Decision SynthesisFinancial Analysis & Trading Decisions98%HIGH14x
Content To Image Prompt GenerationInfrastructure & Utility99%HIGH14x
Model Specific Prompt AdaptationInfrastructure & Utility97%MEDIUM15x
Promotional Message GenerationSocial & Promotional Content97%RANKED21x
Language IdentificationRelevance, Classification & Matching96%MEDIUM24x
Profiled Document Section AnalysisFinancial Analysis & Trading Decisions98%RANKED26x
Persona Policy Eligibility CheckRelevance, Classification & Matching92%MEDIUM26x
Factual Claim RefinementInfrastructure & Utility93%MEDIUM35x
Batch Text TranslationInfrastructure & Utility92%MEDIUM37x
Engagement Opportunity Triage bestRelevance, Classification & Matching100%RANKED37x
Public Response GenerationSocial & Promotional Content93%HIGH42x
Social Post Portfolio SelectionRelevance, Classification & Matching98%HIGH88x
Short Promotional Teaser GenerationSocial & Promotional Content96%RANKED112x

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.