Weekly recap
What changed in LLM pricing — Week of 24–30 August 2026
In the week of 24–30 August 2026, we recorded 45 price changes (21 cuts, 24 increases) and 10 new models across the LLM providers we track. The biggest cut was Nemotron 3.5 Lightning 30B cached input, down 80% ($0.050 → $0.010 per 1M tokens).
The median cut this week was 20%. Figures are from our daily re-check of official provider pricing; see the live tracker for the full history.
45
Price changes
21
Price drops
24
Price increases
10
New models
Price cuts in Week of 24–30 August 2026
Models that got cheaper, biggest cut first. Click a column to sort.
| Nemotron 3.5 Lightning 30B | Cached input | $0.050/1M | $0.010/1M | -80% | 29 August 2026 | Nvidia |
| GLM 5.2 Fast | Cached input | $0.363/1M | $0.145/1M | -60% | 28 August 2026 | Z.ai (Zhipu AI) |
| gemma-4-31b-it | Cached input | $0.100/1M | $0.040/1M | -60% | 30 August 2026 | |
| DeepSeek V3.1 | Cached input | $0.100/1M | $0.060/1M | -40% | 28 August 2026 | DeepSeek |
| GPT OSS 20B | Input | $0.030/1M | $0.020/1M | -33% | 26 August 2026 | OpenAI |
| GPT-5.6 | Output | $30.00/1M | $20.00/1M | -33% | 25 August 2026 | OpenAI |
| GPT-5.6 Sol | Output | $30.00/1M | $20.00/1M | -33% | 25 August 2026 | OpenAI |
| Nemotron 3.5 Lightning 30B | Output | $0.200/1M | $0.150/1M | -25% | 26 August 2026 | Nvidia |
Price increases in Week of 24–30 August 2026
| Nemotron 3.5 Lightning 30B | Cached input | $0.010/1M | $0.050/1M | +400% | 26 August 2026 | Nvidia |
| GPT-5.6 Luna (Bedrock) | Cached input | $0.022/1M | $0.100/1M | +355% | 27 August 2026 | AWS Bedrock |
| GPT-5.6 Luna (Bedrock) | Output | $1.32/1M | $6.00/1M | +355% | 27 August 2026 | AWS Bedrock |
| GPT-5.6 Luna (Bedrock) | Input | $0.220/1M | $1.00/1M | +355% | 27 August 2026 | AWS Bedrock |
| gemma-4-31b-it | Cached input | $0.040/1M | $0.100/1M | +150% | 29 August 2026 | |
| GPT 5.6 Luna Pro | Cache write | $0.125/1M | $0.250/1M | +100% | 29 August 2026 | OpenAI |
| GPT 5.6 Luna Pro | Cached input | $0.010/1M | $0.020/1M | +100% | 29 August 2026 | OpenAI |
| GPT 5.6 Luna Pro | Output | $0.600/1M | $1.20/1M | +100% | 29 August 2026 | OpenAI |
Models launched in Week of 24–30 August 2026
- gemma-4-31b-it· Google$0.070/1M · 29 August 2026
- GPT 5.6 Luna Pro· OpenAI$0.200/1M · 29 August 2026
- GPT 5.6 Terra Pro· OpenAI$2.00/1M · 29 August 2026
- Qwen3.8 Flash· Alibaba$0.150/1M · 29 August 2026
- DeepSeek V3.1· DeepSeek$0.190/1M · 28 August 2026
- GLM-5.2· Mistral AI$1.40/1M · 27 August 2026
- GLM-5.3-Flash· Z.ai (Zhipu AI)$0.075/1M · 27 August 2026
- mistral-medium-3-5· Mistral AI$1.39/1M · 27 August 2026
- GPT OSS 20B· OpenAI$0.020/1M · 26 August 2026
- GLM-5.3· Z.ai (Zhipu AI)$1.40/1M · 25 August 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.