Skip to main content

Knowledge & language Text Model Rankings

Factual accuracy, long-context understanding, language, instruction following, and false-premise detection.

Sources
182.060.031M$8/M$40/M$0.2/M2026-11-09
272.255.641.1M$10/M$50/M2026-09-03
370.766.071.1M$2/M$10/M2026-10-08
467.869.171M$4/M$20/M$0.2/M2026-09-22
566.451.811M$10/M$50/M$0.25/M2026-09-01
665.164.841M$2/M$10/M$0.1/M2026-09-28
764.753.341.1M---2026-07-09
864.561.051.1M$1/M$5/M2026-10-26
964.453.751M-2026-08-03
1064.254.481.1M$2/M$10/M2026-09-22
1163.959.81M-2026-09-30
1263.351.811M$5/M$25/M$0.5/M2026-05-28
1362.845.951.1M$5/M$30/M2026-04-23
1462.444.051M$10/M$50/M$1/M2026-06-09
1561.553.331M$5/M$25/M$0.5/M2026-07-24
1661.062.651M-2026-09-02
1760.758.62500K-2026-09-21
1858.940.021M-2026-05-21
1958.968.661M-2026-09-20
2058.861.261M$3/M$15/M-2026-07-16
2158.748.86500K-2026-08-12
2257.949.11M$1.25/M$4.25/M$0.15/M2026-09-02
2357.264.941M-2026-10-07
2456.779.311M-2026-09-10

Page 1 of 3

Efficiency Frontier

The line shows the best score available at each budget. Models below it have a higher-scoring alternative within the same budget.

Price weighted: cache 7 · input 2 · output 1 (per million tokens)

Frequently asked questions