Weekly recap
What changed in LLM pricing — Week of 5–11 October 2026
In the week of 5–11 October 2026, we recorded 19 price changes (11 cuts, 8 increases) and 14 new models across the LLM providers we track. The biggest cut was Claude Sonnet 5.5 cached input, down 50% ($0.200 → $0.100 per 1M tokens).
The median cut this week was 22%. Figures are from our daily re-check of official provider pricing; see the live tracker for the full history.
19
Price changes
11
Price drops
8
Price increases
14
New models
Price cuts in Week of 5–11 October 2026
Models that got cheaper, biggest cut first. Click a column to sort.
| Claude Sonnet 5.5 | Cached input | $0.200/1M | $0.100/1M | -50% | 8 October 2026 | Anthropic |
| Claude Sonnet 5.5 (Bedrock) | Cached input | $0.200/1M | $0.100/1M | -50% | 8 October 2026 | AWS Bedrock |
| MiniMax-M2.5-highspeed | Cached input | $0.060/1M | $0.030/1M | -50% | 8 October 2026 | MiniMax |
| Nemotron 3.5 Lightning 30B | Input | $0.049/1M | $0.025/1M | -49% | 9 October 2026 | Nvidia |
| Nemotron 3.5 Lightning 30B | Cached input | $0.025/1M | $0.013/1M | -48% | 9 October 2026 | Nvidia |
| Nemotron 3.5 Lightning 30B | Output | $0.180/1M | $0.140/1M | -22% | 8 October 2026 | Nvidia |
| Qwen3-VL 30B-A3B Instruct | Output | $0.520/1M | $0.430/1M | -17% | 10 October 2026 | Alibaba |
| Qwen3-VL 30B-A3B Instruct | Input | $0.130/1M | $0.108/1M | -17% | 10 October 2026 | Alibaba |
Price increases in Week of 5–11 October 2026
| Qwen3 VL 8B Instruct | Input | $0.080/1M | $0.180/1M | +125% | 10 October 2026 | Alibaba |
| GPT OSS 20B | Input | $0.009/1M | $0.018/1M | +100% | 5 October 2026 | OpenAI |
| GPT OSS 20B | Output | $0.045/1M | $0.090/1M | +100% | 5 October 2026 | OpenAI |
| GPT OSS 20B | Cached input | $0.005/1M | $0.009/1M | +80% | 5 October 2026 | OpenAI |
| Qwen3-Vl 30b A3b Thinking | Input | $0.200/1M | $0.290/1M | +45% | 10 October 2026 | Alibaba |
| Qwen3 VL 8B Instruct | Output | $0.500/1M | $0.680/1M | +36% | 10 October 2026 | Alibaba |
| Nemotron 3.5 Lightning 30B | Input | $0.039/1M | $0.049/1M | +26% | 8 October 2026 | Nvidia |
| Nemotron 3.5 Lightning 30B | Cached input | $0.020/1M | $0.025/1M | +25% | 8 October 2026 | Nvidia |
Models launched in Week of 5–11 October 2026
- Qwen3 VL 8B Instruct· Alibaba$0.180/1M · 10 October 2026
- Qwen3-Vl 30b A3b Thinking· Alibaba$0.290/1M · 10 October 2026
- Qwen3-VL 30B-A3B Instruct· Alibaba$0.108/1M · 10 October 2026
- WizardLM-2 8x22B· Microsoft$0.480/1M · 10 October 2026
- Codestral 2· Google$0.300/1M · 9 October 2026
- Nano Banana 2.1· Google$1.50/1M · 9 October 2026
- Nemotron 3.5 Lightning 30B· Nvidia$0.025/1M · 9 October 2026
- Claude Haiku 5.5· Anthropic$0.100/1M · 8 October 2026
- Claude Haiku 5.5 (Bedrock)· AWS Bedrock$0.110/1M · 8 October 2026
- mistral-large-4-0· Mistral AI$0.680/1M · 8 October 2026
- Nemotron 3.5 Lightning 30B· Nvidia$0.049/1M · 8 October 2026
- Mistral Large 4· Mistral AI$0.680/1M · 7 October 2026
- DeepSeek V4.1 Flash Fast· DeepSeek$0.300/1M · 5 October 2026
- GLM 5.3 Fast· Z.ai (Zhipu AI)$1.12/1M · 5 October 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.