Weekly recap
What changed in LLM pricing — Week of 7–13 September 2026
In the week of 7–13 September 2026, we recorded 21 price changes (8 cuts, 13 increases) and 11 new models across the LLM providers we track. The biggest cut was gemma-4-31b-it cached input, down 50% ($0.020 → $0.010 per 1M tokens).
The median cut this week was 19%. Figures are from our daily re-check of official provider pricing; see the live tracker for the full history.
21
Price changes
8
Price drops
13
Price increases
11
New models
Price cuts in Week of 7–13 September 2026
Models that got cheaper, biggest cut first. Click a column to sort.
| gemma-4-31b-it | Cached input | $0.020/1M | $0.010/1M | -50% | 11 September 2026 | |
| Qwen3 Coder Next (Bedrock) | Output | $1.80/1M | $1.20/1M | -33% | 12 September 2026 | AWS Bedrock |
| gemma-3-4b-it | Output | $0.100/1M | $0.080/1M | -20% | 11 September 2026 | |
| gemma-3-4b-it | Input | $0.050/1M | $0.040/1M | -20% | 11 September 2026 | |
| GPT OSS 120B | Input | $0.039/1M | $0.032/1M | -18% | 9 September 2026 | OpenAI |
| gemma-4-31b-it | Output | $0.300/1M | $0.250/1M | -17% | 11 September 2026 | |
| Qwen/Qwen3-Next-80B-A3B-Instruct (Bedrock) | Output | $1.40/1M | $1.20/1M | -14% | 12 September 2026 | AWS Bedrock |
| GLM-5.1 FP8 | Output | $3.30/1M | $3.08/1M | -7% | 10 September 2026 | Z.ai (Zhipu AI) |
Price increases in Week of 7–13 September 2026
| Google Gemma 3 12B (Bedrock) | Output | $0.100/1M | $0.290/1M | +190% | 11 September 2026 | AWS Bedrock |
| Qwen3 Coder Next (Bedrock) | Input | $0.220/1M | $0.500/1M | +127% | 12 September 2026 | AWS Bedrock |
| DeepSeek V4 Flash | Output | $0.280/1M | $0.600/1M | +114% | 11 September 2026 | DeepSeek |
| Google Gemma 3 27B Instruct (Bedrock) | Input | $0.120/1M | $0.230/1M | +92% | 11 September 2026 | AWS Bedrock |
| Google Gemma 3 27B Instruct (Bedrock) | Output | $0.200/1M | $0.380/1M | +90% | 11 September 2026 | AWS Bedrock |
| Google Gemma 3 12B (Bedrock) | Input | $0.050/1M | $0.090/1M | +80% | 11 September 2026 | AWS Bedrock |
| Qwen/Qwen3-VL-235B-A22B-Instruct (Bedrock) | Output | $1.50/1M | $2.66/1M | +77% | 12 September 2026 | AWS Bedrock |
| Qwen/Qwen3-VL-235B-A22B-Instruct (Bedrock) | Input | $0.300/1M | $0.530/1M | +77% | 12 September 2026 | AWS Bedrock |
Models launched in Week of 7–13 September 2026
- Hy-MT2 Plus· Tencent$0.074/1M · 13 September 2026
- DeepSeek V4.1 Flash· DeepSeek$0.150/1M · 11 September 2026
- deepseek-v4-1-flash· DeepSeek$0.150/1M · 11 September 2026
- Gemma 4 26B A4B IT (Bedrock)· AWS Bedrock$0.130/1M · 11 September 2026
- Gemma 4 31B IT (Bedrock)· AWS Bedrock$0.140/1M · 11 September 2026
- Gemma 4 E2B IT (Bedrock)· AWS Bedrock$0.040/1M · 11 September 2026
- gemma-3-4b-it· Google$0.040/1M · 11 September 2026
- GPT-5.6 Luna (India) (Bedrock)· AWS Bedrock$0.220/1M · 11 September 2026
- GPT-5.6 Terra (India) (Bedrock)· AWS Bedrock$2.20/1M · 11 September 2026
- GLM-5.1 FP8· Z.ai (Zhipu AI)$0.980/1M · 10 September 2026
- GPT-6 Astra (Bedrock)· AWS Bedrock$11.00/1M · 9 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.