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

Claude Sonnet 4.6 at a glance

Good enough on 19/52 tasks at the 90% bar. Best value on 0 tasks. Best fit for: Long-form Content Generation.

Provider
Anthropic
Model name
claude-sonnet-4-6
Cost mode:
Qualifies on
19 / 52 tasks (at 90% bar)
Best value on
0 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)Theme Generation — quality 8.77, 10.8x the cost of the best-value good-enough optionActivity Feed Blurb Generation — quality 8.19, 18.3x the cost of the best-value good-enough optionClaim-Referenced Analyst Writing — quality 8.78, 5.2x the cost of the best-value good-enough optionInvestment Panel Voting — quality 9.09, 5.1x the cost of the best-value good-enough optionGeographic Region Identification — quality 9.10, 15.5x the cost of the best-value good-enough optionSocial Post Promotion — quality 8.02, 9.9x the cost of the best-value good-enough optionOnboarding Chapter Prompt Adaptation — quality 8.66, 40.7x the cost of the best-value good-enough optionTrading Recommendation — quality 8.59, 19.2x the cost of the best-value good-enough optionLanguage Detection — quality 9.96, 51.7x the cost of the best-value good-enough optionExecutive Summary Generation — quality 7.81, 29.8x the cost of the best-value good-enough optionPublication Title Generation — quality 8.25, 24.2x the cost of the best-value good-enough optionTopic Cluster Naming — quality 8.60, 6.7x the cost of the best-value good-enough optionEngagement Triage — quality 8.92, 34.5x the cost of the best-value good-enough optionPrompt Adaptation — quality 8.90, 13.5x the cost of the best-value good-enough optionX Post Selection — quality 8.71, 22.8x the cost of the best-value good-enough optionAuthor Voice Generation — quality 9.12, 58.1x the cost of the best-value good-enough optionAuthor Matching — quality 8.21, 3.6x the cost of the best-value good-enough optionImage Prompt Generation — quality 8.31, 50.7x the cost of the best-value good-enough optionTopic-to-Section Assignment — quality 8.12, 12.8x 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
Author MatchingRelevance, Classification & Matching91%MEDIUM3.6x
Investment Panel Voting bestFinancial Analysis & Trading Decisions100%RANKED5.1x
Claim-Referenced Analyst WritingLong-form Content Generation90%HIGH5.2x
Topic Cluster NamingTopic Organization & Clustering98%HIGH6.7x
Social Post PromotionSocial & Promotional Content90%HIGH9.9x
Theme GenerationLong-form Content Generation90%HIGH11x
Topic-to-Section AssignmentTopic Organization & Clustering90%HIGH13x
Prompt Adaptation bestInfrastructure & Utility100%RANKED13x
Geographic Region IdentificationStructured Data & Fact Extraction97%HIGH16x
Activity Feed Blurb GenerationSocial & Promotional Content93%MEDIUM18x
Trading RecommendationFinancial Analysis & Trading Decisions99%HIGH19x
X Post SelectionRelevance, Classification & Matching100%HIGH23x
Publication Title GenerationContent Summarization & Synthesis92%RANKED24x
Executive Summary GenerationContent Summarization & Synthesis98%MEDIUM30x
Engagement Triage bestRelevance, Classification & Matching100%RANKED34x
Onboarding Chapter Prompt AdaptationInfrastructure & Utility96%HIGH41x
Image Prompt GenerationInfrastructure & Utility90%RANKED51x
Language DetectionRelevance, Classification & Matching99%RANKED52x
Author Voice GenerationLong-form Content Generation94%RANKED58x

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.