Best LLM for…

The best cheap LLMs (under $0.15 per million tokens)

Answer: Mercury 2.5 by Inception: 57.5% at $0.00011 per question.

Ranking

Models with a blended list price under $0.15 per million tokens (3 input : 1 output), ordered by verification score.

The best cheap LLMs (under $0.15 per million tokens)
#ModelScoreCost / questionCost / correct answerSpeed (tok/s)
1Mercury 2.5Inception57.5%$0.00011$0.00019157
2gpt-oss-20bOpenAI51.2%$0.000066$0.00013103
3Nemotron 3.5 LightningNVIDIA43.8%$0.00049$0.0011116
4Ling 3.1 FlashinclusionAI41.2%$0.00061$0.00157
5Phi 4Microsoft41.2%$0.00012$0.00029103
6Solar Mini 4Upstage41.2%$0.00053$0.0013107
7Granite 4.2 8BIBM36.2%$0.000077$0.0002160
8Ministral 3 3B 2512Mistral26.2%$0.00019$0.00074252
9Granite 4.0 MicroIBM25.0%$0.000051$0.000238
10Nova Micro 1.0Amazon22.5%$0.000087$0.00039180
11Gemma 4 26B A4BGoogle15.0%$0.0011$0.0072172
12Reka Flash 3rekaai1.2%$0.0008$0.067101

How to read this

Scores come from 80 multiple-choice questions from SuperGPQA and MMLU-Pro. With that many questions, a score has a margin of error of several points, so models a few points apart are effectively tied on accuracy; the comparison pages run a significance test for any pair. To see the top two side by side, compare Mercury 2.5 and gpt-oss-20b.

Speed, latency and cost are measured through OpenRouter and depend on which host serves the request; your numbers on a direct API may differ. The full procedure is on the methodology page, and the data behind this ranking can be downloaded.

Other questions

Quick answers

What is the best low-cost AI model?
Mercury 2.5 by Inception: 57.5% at $0.00011 per question. Rule: Models with a blended list price under $0.15 per million tokens (3 input : 1 output), ordered by verification score.
What is the runner-up?
gpt-oss-20b by OpenAI: 51.2% at $0.000066 per question.
Which of these costs least per correct answer?
gpt-oss-20b: $0.00013 per correct answer (cost per question divided by its score).