Model Catalog
Live model listing from OpenRouter β filter, search, and sort
| β | Model Name β | Provider β | Flag | Input $/M β | Output $/M β | Context β | Modality |
|---|---|---|---|---|---|---|---|
| Loading models⦠| |||||||
𧡠Parallel Coordinates
Each line = one model. Check boxes in the table above to highlight models for comparison.
Find Your Perfect AI Model
Answer 4 questions β get objective, data-driven recommendations.
What's your primary use case?
What's your budget?
Any specific requirements? (select all that apply)
What matters most to you?
π― Your Top Recommendations
Intelligence
Price tracking, comparisons, momentum, and value analysis
What changed this week across OpenRouter's model catalog
π₯ Most Volatile
π§ Most Stable (never changed)
π Currently at All-Time Low (models with price changes in last 2 weeks)
π Price Rank Changes
Rank 1 = cheapest. Line connects 30d ago β 7d ago β Today. Rising line = became relatively cheaper.
π Price History
βΉ Prices approximate as of Feb 2026 β verify at provider website. Source links provided where available.
| Model β | Provider β | OR Input $/M β β | Direct Input $/M β | OR Output $/M β β | Direct Output $/M β | Savings % β | Source |
|---|---|---|---|---|---|---|---|
| Loading direct pricing data⦠| |||||||
HuggingFace download trends over the last 7 days β open-weight models only
βΉ Only open-weight models with HuggingFace data appear here. Proprietary models (OpenAI, Anthropic, Google) are excluded β they lack HF download statistics. X-axis: input $/M tokens (log scale) Β· Y-axis: HF quality proxy (downloads + likes) Β· Bubble size: context length.
β EU provider
β US/Canada
β Other
β China-based
Free Models Leaderboard
Ranked by composite score β adjust weights to re-rank in real time
Privacy Rankings
Scored on open weights, license, provider geography, and moderation
0
Best for regulated industries
Loadingβ¦
| # | Model | Provider | Country | Privacy Score β | License | Open Weights | Pricing |
|---|
Benchmark Scores
HF Open LLM Leaderboard 2 β open models only. Click column headers to sort.
βΉ Note: benchmark data only covers open models listed on the HuggingFace Open LLM Leaderboard 2. Proprietary models (OpenAI, Anthropic, Google) are not included.
Highlight:
| Model | Provider | Avg β | Arena ELO β | IFEval β | BBH β | MATH-Hard β | GPQA β | MMLU-Pro β | MuSR β | Bench Efficiency | Price/M in β | Price/M out β | Peers Ξ β |
|---|
How "Peers Ξ" is computed:
For each paid model with benchmark data, we locate its k=5 nearest neighbors in the 6-dimensional benchmark space (IFEval, BBH, MATH-Hard, GPQA, MMLU-Pro, MuSR). Each dimension is z-score standardized (mean=0, std=1) so no single benchmark dominates. When a model is missing some benchmark values, distance is computed over shared dimensions only and scaled to a 6-dimension equivalent. The fair price is the inverse-distance-weighted (IDW) average of those 5 peers' prices β closer peers have higher weight. The signed percentage shows how the model's actual price deviates: negative % = underpriced relative to peers (good value), positive % = overpriced, Β±15% = within fair range. Hover any Peers Ξ cell to see the exact fair-price estimate and the 5 peer models. Only covers paid models listed on the HF Open LLM Leaderboard 2.
Provider Intelligence
Model counts, pricing, and geographic breakdown
Model count by provider
Click a cell to see that provider's price breakdown β
Input vs output cost
Distribution of input $/M tokens per provider (log scale) Β· click treemap to focus a provider
Geographic breakdown & regulatory notes
Provider headquarters
Dot size = model count Β· click markers for details
Provider directory
| Provider | HQ | Founded | Key founders | Latest funding | Links | Wiki views |
|---|---|---|---|---|---|---|
| Loading⦠| ||||||