AI IQ
Subscribe

ProofBench Cost Efficiency

Archived chart. Historical results combine benchmark versions and are excluded from IQ. Use the separately versioned ProofBench 1.1 for current comparisons.

ProofBench Cost Efficiency
ProofBench Cost Efficiency
Archived chart. Historical results combine benchmark versions and are excluded from IQ. Use the separately versioned ProofBench 1.1 for current comparisons.

How to read this chart

ProofBench publishes no per-run cost, so this chart falls back to each model's estimated effective cost per 1M I/O tokens (a runtime estimate from measured or imputed token usage, not a published task cost). Explicit reasoning/nonreasoning siblings are connected where both are public; every other model is a standalone point. Up and to the left is better.

Data sources
Artificial AnalysisARC PrizeVals Index v2 Cost per TestVals.ai
Archived chartProofBench vs Effective CostArchived chart. Historical results combine benchmark versions and are excluded from IQ. Use the separately versioned ProofBench 1.1 for current comparisons.Data: Artificial Analysis, ARC Prize, Vals Index v2 Cost per Test +1 moreOpen chartArchived chartProofBench Benchmark ScoresArchived chart. Historical results combine benchmark versions and are excluded from IQ. Use the separately versioned ProofBench 1.1 for current comparisons.Data: Vals.aiOpen chartArchived chartFrontierSWE ScoresArchived chart. V1 Dominance retired from Composite IQ and Software Engineering; V2 mean@5 requires a separate calibration.Data: FrontierSWEOpen chartMore / ExperimentalSWE-Bench Verified Benchmark ScoresHistorical data. Vals archived cohort; retired from default Software Engineering domain. Each model's SWE-Bench Verified score. Color = provider.Data: Vals.ai, LLM StatsOpen chartArchived chartIOI Benchmark ScoresArchived chart. IOI V1 archived; V2 uses a different harness and 2024–2026 tasks.Data: Vals.ai IOIOpen chartMore / ExperimentalTerminal-Bench 2.0 Benchmark ScoresHistorical data. Superseded by 2.1 and 4.0. Legacy Terminal-Bench 2.0 scores retained for historical comparison. Color = provider.Data: Terminal-BenchOpen chart