Bertrande

Cheapest model that is still capable

What is the cheapest model above our capability threshold?

As of 18 September 2026, the model ranking first for this question is Upstage: Solar Pro 4 on OpenRouter, with intelligence index 41.6 and an output price of $0.360 per million tokens. Ranked by lowest output price among models scoring at least 40 on intelligence, across 22 models for which we hold both figures. This ranking ignores speed and reliability: it is a query on price and a relayed capability score, not a verdict on which model serves you best.

How this ranking is built

Criterion: lowest output price among models scoring at least 40 on intelligence. The threshold of 40 is our stated cut-off, not a measured boundary. We say it plainly so you can disagree with it and re-rank the table yourself. Every model below the line is still in the catalogue.

This is a query, not an opinion. Both figures are shown on every row so you can check the order, and the table sorts on any column if you disagree with our criterion.

Two things to know before trusting this table. Capability scores are published by Artificial Analysis and relayed by the catalogue we read. They are not our measurement, and a score measures the model rather than the seller, so we apply it to every seller of that model. This ranking says nothing about speed or reliability, which we measure only in part: availability is published per upstream host in reliability.json and on the model pages, and we publish no latency at all, for reasons given on the data page. Prices are ours to observe, each carries the date we saw it, and we keep only the cheapest offer per model so that batch and free tiers do not crowd out the answer.

#ModelSellerIntelligence index/ 10022 of 22 heldScale from 3.8 to 55Input$ / 1M tokens22 of 22 heldScale from $0.0001 to $600Output$ / 1M tokens22 of 22 heldScale from $0.0001 to $600Contexttokens14 of 22 heldScale from 512 to 2MObserved
1Upstage: Solar Pro 4OpenRouter41.6$0.090$0.360524 28818 September 2026
2GLM-5.3-FlashZ.ai first party41.9$0.150$0.500n/a18 September 2026
3Nex AGI: Nex-N2-ProOpenRouter41.7$0.250$1.00262 1448 September 2026
4Qwen: Qwen3.6 PlusOpenRouter40.5$0.325$1.951 000 00018 September 2026
5grok-build-0.1xAI first party40.7$1.00$2.00256 00018 September 2026
6Gemini 3.8 FlashGoogle first party41.2$0.750$3.75n/a18 September 2026
7MoonshotAI: Kimi K2.7 Code (batch)OpenRouter43.0$0.950$4.00262 14427 August 2026
8GLM-5.3Z.ai first party44.9$1.40$4.40n/a18 September 2026
9Qwen/Qwen3.8-MaxDeepInfra53.4$1.65$4.951256 00015 September 2026
10Qwen: Qwen3.8 Max (0902)OpenRouter45.4$2.00$6.001 000 00016 September 2026
11grok-4.6xAI first party44.4$2.00$6.00500 00018 September 2026
12Qwen/Qwen3.8-2.4T-A95BDeepInfra40.0$2.00$6.00262 14418 September 2026
13gpt-5.6-terraOpenAI first party42.3$2.00$12.00n/a18 September 2026
14OpenAI: GPT-5.4OpenRouter53.1$2.50$15.001 050 00018 September 2026
15kimi-k3Moonshot AI first party43.8$3.00$15.001 048 57618 September 2026
16gpt-5.6-solOpenAI first party47.1$4.00$20.00n/a18 September 2026
17Claude Opus 4.7Anthropic first party55.0$5.00$25.00n/a18 September 2026
18Claude Opus 5Anthropic first party50.7$5.00$25.001 000 00018 September 2026
19Claude Opus 4.8Anthropic first party42.0$5.00$25.00n/a18 September 2026
20Claude Fable 5.1Anthropic first party53.4$10.00$50.001 000 00018 September 2026
21gpt-6-astraOpenAI first party52.8$10.00$50.00n/a18 September 2026
22Claude Fable 5Anthropic first party49.7$10.00$50.001 000 00018 September 2026

All 22 models meeting this criterion are listed.

A model missing from this table is missing one of the two figures, not judged. We would rather show a shorter list than fill it with guesses.

Other questions we can answer this way