LLM comparison
The major models side by side: price, context window, size, and whether you can actually own the weights. Click any column to sort. Want two head to head? See thehead-to-head comparisonsor the always-current LLM index (with JSON download). The cheapest per-token model is not always the cheapest per finished task, so pair this with thetoken counter.
| Model | Weights | Params (B) | Context | $/M in | $/M out |
|---|
Approximate public figures as of mid-2026; prices change often. Verify live provider pricing before relying on these numbers.
Open vs closed is the decision under the decision
Price and context get the attention, but the column that compounds is open vs closed. An open model you can download, fine-tune, run offline, and pin to a version, which is why a release likeGLM-5.2 matters beyond its sticker price. If you are weighing running a model yourself versus paying per token, theself-host vs API calculator turns it into a break-even number, and the VRAM calculator tells you what hardware it takes.