Best LLMs Overall
A single general-purpose ranking hides more than it reveals, which is why this site also maintains per-task pages for coding, research, and specific capabilities. Still, a defensible generalist ordering is useful as a starting point, and this page provides one: models ranked by composite intelligence-index evidence, shown with current pricing so the cost of each quality tier is visible.
The practical takeaway from this table is usually not "use the number one model for everything" but the size of the price gap between adjacent quality tiers. That gap is the entire economic case for routing requests to the model each one actually needs.
The top pick
The strongest measured fit for this task right now, taken from the live ranking below.
- Intelligence Index
- 57.6
- Context
- 1M
The ranking
Ordered from live catalog data; the note beneath the table states exactly how. Missing values are shown as missing, never estimated.
| # | Model | Quality index | Intelligence Index | Context |
|---|---|---|---|---|
| 1 | anthropic | 100 | 57.6 | 1M |
| 2 | anthropic | 99.5 | 56.0 | 1M |
| 3 | anthropic | 99 | 53.4 | 1M |
| 4 | openai | 98.6 | 52.7 | 1.1M |
| 5 | google | 98.1 | 52.6 | 1M |
| 6 | openai | 97.6 | 51.8 | 1.1M |
| 7 | anthropic | 97.1 | 50.8 | 1M |
| 8 | anthropic | 96.7 | 49.6 | 1M |
| 9 | openai | 96.2 | 47.6 | 1.1M |
| 10 | openai | 95.7 | 47.0 | 1.1M |
This list is ordered by the same per-task quality projection the Inferbase router uses to route production traffic: benchmark results from independent suites are normalized and combined per task. The quality index is a 0-100 percentile among the 199 models with measured evidence for this task. Models without published benchmark results here are not ranked, and not guessed at.
marks models served through the Inferbase API. Missing values are shown as No data rather than estimated.
Also worth evaluating
The next ranks after the top ten, for teams that want a wider shortlist.
Or stop choosing manually
The honest answer changes as benchmarks and prices move. Routing makes the choice per request instead of once.
Send model="auto" to one OpenAI-compatible endpoint and each request is served by the best fit from the same quality and price data behind this ranking, with the decision disclosed per request.
Every routed request records the task it was classified as, the models that qualified, the one that won and why, so a pick made here can be checked against what the router chose in production. A model pinned by name is served as named, with no routing decision.
from openai import OpenAI
client = OpenAI(
base_url="https://api.inferbase.ai/v1",
api_key="YOUR_INFERBASE_KEY",
)
# Let routing pick the best model per request,
# or pin any model id from GET /models.
response = client.chat.completions.create(
model="auto",
messages=[{"role": "user", "content": "..."}],
)Frequently asked questions
How this ranking works and how to act on it.
It is a composite of many individual evaluations published by an independent benchmark suite, normalized across the catalog. It is the generalist signal used when a task has no dedicated benchmark, both on this page and inside the Inferbase router.
Because most production requests are not frontier-difficulty. Classification, extraction, short summaries, and routine Q&A are served indistinguishably well by models far down this table at a fraction of the price. The top model is the right tool for the hardest slice of traffic, not all of it.
The gap has narrowed substantially: current open-weight models place high in this ranking while costing dramatically less per token, and they can be self-hosted. The table makes the current state of that comparison concrete rather than anecdotal.
From independent benchmark suites ingested into the Inferbase catalog and projected into per-task quality scores, the same data path the production router uses. Creator self-reported numbers are not the ranking basis.
It regenerates at least hourly from live catalog data, and materially reorders when a new model publishes benchmark results or a price changes. The month in the title tracks the data, not a manual rewrite.
The model at the top of this table on the date shown, by composite intelligence evidence across the independent suites in the catalog. The order changes as new models are measured and benchmark rounds are published, which is why the page carries its refresh date and the ranking is recomputed from live catalog data rather than written once.
Read the price gap between adjacent rows. The quality index moves in small steps down the table while the price per token often drops by a multiple, so the best value is the cheapest model whose index is within the margin your task tolerates. For a specific task, the per-task pages rank on that task’s evidence, and the recommender turns a use case into a shortlist.
On this page, on the same index: every model with published results on the general suites is ranked here regardless of creator, proprietary or open-weight. Open any two for a side-by-side comparison of their specifications, capabilities and benchmark scores by suite, with a cell shown as not available where a suite has not measured a model.
Related rankings
The same evidence-first method, applied to other tasks.
Put intelligence in the middle.
One gateway in front of every model, with your policies applied and every decision on record. Start on the free plan with 5,000 requests a month, no card required.