AgMoDB
ModelsAgentsOrchestratorsEvalsCompositesVisualizeIndustry
AgMoDB by @mistakeknot
All Benchmarks

Parameter Count

Total number of trainable parameters in the model, as catalogued by Epoch AI's frontier model database.

aggregateepoch_ai

Strengths

Simple, well-understood metric that correlates with model capability. Consistently reported across most models.

Caveats

Parameter count alone doesn't determine performance — architecture, training data quality, and compute all matter. MoE models complicate comparisons.

How to interpret scores

Contamination risk: moderateFreshness: periodic

Lower is better for deployment efficiency at similar quality. Values are parameter counts and can range from millions to trillions.

Relevant use cases

Model sizingHardware requirementsArchitecture comparison

Caveat to keep in mind

Parameter count alone doesn't determine performance — architecture, training data quality, and compute all matter.

Leaderboard

Last synced: Nov 22, 2026, 4:18 AM
#ModelProviderScorePrice/M
1
PALM-2
source epoch_ai
Evidence observed Nov 22, 2026
Google340000000000.0$0.00
2
Meta: Llama 3.1 405B (base)
source epoch_ai
Evidence observed Nov 22, 2026
Meta405000000000.0$4.00
3
Meta: Llama 3.1 405B Instruct
source epoch_ai
Evidence observed Mar 6, 2026
Meta405000000000.0$4.00
4
Writer: Palmyra X5
source epoch_ai
Evidence observed Mar 6, 2026
Writer540350000000.0$1.95
5
NVIDIA: Switchyard
source epoch_ai
Evidence observed Nov 22, 2026
NVIDIA1571000000000.0$-1000000.00
6
Switchpoint Router
source epoch_ai
Evidence observed Oct 2, 2026
Switchpoint1571000000000.0$1.49
7
GPT-4
source epoch_ai
Evidence observed Nov 22, 2026
OpenAI1800000000000.0$37.50
8
Grok 4
source epoch_ai
Evidence observed Nov 22, 2026
SpaceXAI3000000000000.0$6.00
9
Grok 3
source epoch_ai
Evidence observed Nov 22, 2026
SpaceXAI3000000000000.0$6.00