GPT-5.6 Luna gained 3.3 points of share
From 10.5% to 13.9% between the last two 7-day windows.
What the data says. OpenAI has 18.6% of tokens and 24.9% of estimated spend this week.
See GPT-5.6 Luna on Model Season →GPT-5.6 Luna gained 3.3 points of share
GLM 5.3 Flash got 40% cheaper
From $0.24 to $0.14 per 1M tokens, blended. Input $0.150 → $0.090, output $0.50 → $0.30. New price since Sep 16.
What the data says. The model has 9.02% of traffic over the last 7 days, ranked 3rd. Z.ai (GLM) has 13.6% of tokens and 8.5% of estimated spend this week.
See GLM 5.3 Flash on Model Season →GLM 5.3 entered the top 10
Ranked 10th, with 2.0% of traffic over the last 7 days.
What the data says. Z.ai (GLM) has 13.6% of tokens and 8.5% of estimated spend this week.
See GLM 5.3 on Model Season →Union Alpha is now in the catalog
Anonymous (stealth), $0.0000 input and $0.0000 output per 1M tokens and 256k context. No traffic volume measured yet.
What the data says. The model has 0.11% of traffic over the last 7 days, ranked 51st.
See Union Alpha on Model Season →Union Alpha debuted in traffic
First volume recorded on Sep 16, 2026, 0 days after entering the catalog, with 0.11% of traffic over the last 7 days.
See Union Alpha on Model Season →Other changes
- PriceNemotron 3 Ultra got 19% more expensive. From $1.05 to $1.25 per 1M tokens, blended. Input $0.60 → $0.63, output $2.40 → $3.13. New price since Sep 16. See model →
- PriceMythoMax 13B got 46% more expensive. From $0.060 to $0.087 per 1M tokens, blended. Input $0.060 → $0.080, output $0.060 → $0.110. New price since Sep 16. See model →
- PriceMistral Small 3.2 24B got 25% more expensive. From $0.11 to $0.13 per 1M tokens, blended. Input $0.075 → $0.094, output $0.20 → $0.25. New price since Sep 16. See model →
- PriceMiniMax M1 got 12% cheaper. From $0.96 to $0.85 per 1M tokens, blended. Input $0.55 → $0.40, output $2.20 → $2.20. New price since Sep 16.
Catalog snapshot of Sep 17, 2026, compared with the previous ones since Sep 12, 2026. Traffic through Sep 16, 2026. Price is a 75% input and 25% output blend, and only becomes news when the new value holds for two snapshots. Order follows a fixed relevance formula that weighs the type of change and the model’s share.