Weekly recap
What changed in LLM pricing — Week of 14–20 September 2026
In the week of 14–20 September 2026, we recorded 19 price changes (12 cuts, 7 increases) and 9 new models across the LLM providers we track. The biggest cut was Mistral Small 3.1 24B Instruct input, down 72% ($0.351 → $0.100 per 1M tokens).
The median cut this week was 18%. Figures are from our daily re-check of official provider pricing; see the live tracker for the full history.
19
Price changes
12
Price drops
7
Price increases
9
New models
Price cuts in Week of 14–20 September 2026
Models that got cheaper, biggest cut first. Click a column to sort.
| Mistral Small 3.1 24B Instruct | Input | $0.351/1M | $0.100/1M | -72% | 15 September 2026 | Mistral AI |
| Qwen3.8 2.4T A95B | Cached input | $0.500/1M | $0.200/1M | -60% | 17 September 2026 | Alibaba |
| Mistral Small 3.1 24B Instruct | Output | $0.555/1M | $0.300/1M | -46% | 15 September 2026 | Mistral AI |
| deepseek-v4-1-flash | Output | $0.600/1M | $0.400/1M | -33% | 15 September 2026 | DeepSeek |
| deepseek-v4-1-flash | Input | $0.150/1M | $0.100/1M | -33% | 15 September 2026 | DeepSeek |
| Qwen3.8 27B | Input | $0.100/1M | $0.080/1M | -20% | 17 September 2026 | Alibaba |
| NVIDIA Nemotron 3 Ultra | Cached input | $0.120/1M | $0.100/1M | -17% | 15 September 2026 | Nvidia |
| NVIDIA Nemotron 3 Ultra | Input | $0.600/1M | $0.500/1M | -17% | 15 September 2026 | Nvidia |
Price increases in Week of 14–20 September 2026
| mistral-medium-3-5 | Cached input | $0.139/1M | $0.750/1M | +440% | 15 September 2026 | Mistral AI |
| Qwen3.8 27B | Cached input | $0.010/1M | $0.050/1M | +400% | 17 September 2026 | Alibaba |
| Qwen3.8 2.4T A95B | Cached input | $0.250/1M | $0.500/1M | +100% | 15 September 2026 | Alibaba |
| Mistral Small 3.2 24B Instruct | Input | $0.075/1M | $0.094/1M | +25% | 16 September 2026 | Mistral AI |
| Mistral Small 3.2 24B Instruct | Output | $0.200/1M | $0.250/1M | +25% | 16 September 2026 | Mistral AI |
| mistral-medium-3-5 | Input | $1.39/1M | $1.50/1M | +8% | 15 September 2026 | Mistral AI |
| mistral-medium-3-5 | Output | $7.13/1M | $7.50/1M | +5% | 15 September 2026 | Mistral AI |
Models launched in Week of 14–20 September 2026
- GLM-5.3· Mistral AI$1.40/1M · 18 September 2026
- Qwen3.8 Omni Flash· Alibaba$0.150/1M · 18 September 2026
- meta-llama-3-3-70b-instruct· Meta$0.590/1M · 17 September 2026
- Qwen3.8 2.4T A95B· Alibaba$1.95/1M · 17 September 2026
- Qwen3.8 27B· Alibaba$0.080/1M · 17 September 2026
- deepseek-v4-1-flash· DeepSeek$0.100/1M · 15 September 2026
- Mistral Small 3.1 24B Instruct· Mistral AI$0.100/1M · 15 September 2026
- mistral-medium-3-5· Mistral AI$1.50/1M · 15 September 2026
- NVIDIA Nemotron 3 Ultra· Nvidia$0.500/1M · 15 September 2026
Cite this recap
Monthly recaps are generated from our open price-change dataset (CC BY 4.0). See the recap JSON, the full change dataset, or the RSS feed.