Weekly recap
What changed in LLM pricing — Week of 3–9 August 2026
In the week of 3–9 August 2026, we recorded 12 price changes (10 cuts, 2 increases) and 91 new models across the LLM providers we track. The biggest cut was GPT-5.6 Luna input, down 80% ($1.00 → $0.200 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.
12
Price changes
10
Price drops
2
Price increases
91
New models
Price cuts in Week of 3–9 August 2026
Models that got cheaper, biggest cut first. Click a column to sort.
| GPT-5.6 Luna | Input | $1.00/1M | $0.200/1M | -80% | 9 August 2026 | OpenAI |
| GPT-5.6 Luna | Output | $6.00/1M | $1.20/1M | -80% | 9 August 2026 | OpenAI |
| GPT-5.6 Luna (Bedrock) | Input | $1.00/1M | $0.220/1M | -78% | 9 August 2026 | AWS Bedrock |
| GPT-5.6 Luna (Bedrock) | Output | $6.00/1M | $1.32/1M | -78% | 9 August 2026 | AWS Bedrock |
| Llama-3.3-70B-Instruct | Input | $0.130/1M | $0.100/1M | -23% | 9 August 2026 | Meta |
| GPT-5.6 Terra | Input | $2.50/1M | $2.00/1M | -20% | 9 August 2026 | OpenAI |
| GPT-5.6 Terra | Output | $15.00/1M | $12.00/1M | -20% | 9 August 2026 | OpenAI |
| Llama-3.3-70B-Instruct | Output | $0.400/1M | $0.320/1M | -20% | 9 August 2026 | Meta |
Price increases in Week of 3–9 August 2026
| GPT-5.6 Sol (Bedrock) | Input | $5.00/1M | $5.50/1M | +10% | 9 August 2026 | AWS Bedrock |
| GPT-5.6 Sol (Bedrock) | Output | $30.00/1M | $33.00/1M | +10% | 9 August 2026 | AWS Bedrock |
Models launched in Week of 3–9 August 2026
- Claude 3 Haiku· Anthropic$0.250/1M · 9 August 2026
- Claude 3.5 Haiku· Anthropic$0.800/1M · 9 August 2026
- Claude Opus 4· Anthropic$13.00/1M · 9 August 2026
- Claude Opus 4.8 Fast· Anthropic$10.00/1M · 9 August 2026
- Claude Opus 5· Anthropic$5.00/1M · 9 August 2026
- Claude Opus 5 (Bedrock)· AWS Bedrock$5.00/1M · 9 August 2026
- Claude Opus 5 Fast· Anthropic$10.00/1M · 9 August 2026
- Claude Sonnet 4· Anthropic$2.60/1M · 9 August 2026
- DeepSeek Chat V3.1· DeepSeek$0.200/1M · 9 August 2026
- DeepSeek R1· DeepSeek$0.400/1M · 9 August 2026
- DeepSeek R1 Distill Llama 70B· DeepSeek$0.030/1M · 9 August 2026
- Deepseek R1 Distill Qwen 32B· DeepSeek$0.300/1M · 9 August 2026
- DeepSeek V3· DeepSeek$0.250/1M · 9 August 2026
- DeepSeek V3.1· DeepSeek$0.200/1M · 9 August 2026
- DeepSeek V3.1 Terminus· DeepSeek$0.250/1M · 9 August 2026
- Gemini 3 Flash· Google$0.400/1M · 9 August 2026
- Gemini 3 Pro· Google$1.60/1M · 9 August 2026
- Gemini 3.1 Pro Preview· Google$2.00/1M · 9 August 2026
- Gemini 3.5 Flash Lite· Google$0.300/1M · 9 August 2026
- Gemini 3.6 Flash· Google$1.50/1M · 9 August 2026
- Gemma 3 12B IT· Google$0.050/1M · 9 August 2026
- Gemma 3 27B IT· Google$0.080/1M · 9 August 2026
- Gemma 4 26B A4B IT· Google$0.060/1M · 9 August 2026
- Gemma 4 31B IT· Google$0.080/1M · 9 August 2026
- Gemma 4 E4B IT· Google$0.020/1M · 9 August 2026
- gemma-3-4b-it· Google$0.050/1M · 9 August 2026
- GLM 5.2 Fast· Z.ai (Zhipu AI)$1.45/1M · 9 August 2026
- glm-5-2-fp8· Z.ai (Zhipu AI)$0.850/1M · 9 August 2026
- GLM-5.1 FP8· Z.ai (Zhipu AI)$0.850/1M · 9 August 2026
- GPT 5.6 Luna Pro· OpenAI$0.100/1M · 9 August 2026
- GPT 5.6 Sol Pro· OpenAI$5.00/1M · 9 August 2026
- GPT 5.6 Terra Pro· OpenAI$1.00/1M · 9 August 2026
- GPT OSS 120B· OpenAI$0.032/1M · 9 August 2026
- GPT OSS 20B· OpenAI$0.030/1M · 9 August 2026
- GPT-3.5 Turbo Instruct· OpenAI$1.40/1M · 9 August 2026
- Grok 4· xAI$3.00/1M · 9 August 2026
- Grok 4.20· xAI$1.25/1M · 9 August 2026
- Grok Code Fast 1· xAI$0.180/1M · 9 August 2026
- Kimi K2· Moonshot AI$0.400/1M · 9 August 2026
- Kimi K2 Instruct· Moonshot AI$0.400/1M · 9 August 2026
- Kimi K3 Fast· Moonshot AI$3.00/1M · 9 August 2026
- Llama 3.1 70B Instruct· Meta$0.400/1M · 9 August 2026
- Llama 3.1 8b Instruct· Meta$0.025/1M · 9 August 2026
- Llama 3.2 11B Vision Instruct· Meta$0.055/1M · 9 August 2026
- Llama 3.2 1B Instruct· Meta$0.010/1M · 9 August 2026
- Llama 3.2 3B Instruct· Meta$0.020/1M · 9 August 2026
- Llama 4 Scout 17B 16E Instruct· Meta$0.100/1M · 9 August 2026
- MiniMax 01· MiniMax$0.139/1M · 9 August 2026
- MiniMax M1· MiniMax$0.132/1M · 9 August 2026
- MiniMax M1 80K· MiniMax$0.550/1M · 9 August 2026
- MiniMax M2-her· MiniMax$0.300/1M · 9 August 2026
- Ministral 3B· Mistral AI$0.040/1M · 9 August 2026
- Mistral 7B Instruct v0.3· Mistral AI$0.110/1M · 9 August 2026
- Mistral Large· Mistral AI$2.00/1M · 9 August 2026
- Mistral Medium 3· Mistral AI$0.400/1M · 9 August 2026
- Mistral Medium 3.5 128B· Mistral AI$1.65/1M · 9 August 2026
- Mistral Saba· Mistral AI$0.199/1M · 9 August 2026
- Mistral Small 3.1 24B Instruct· Mistral AI$0.351/1M · 9 August 2026
- Mistral Small 3.2 24B Instruct· Mistral AI$0.094/1M · 9 August 2026
- mistral-medium-3-5· Mistral AI$1.67/1M · 9 August 2026
- Muse Spark 1.2· Meta$1.25/1M · 9 August 2026
- Muse Spark 1.2 Contributor· Meta$0.100/1M · 9 August 2026
- Nemotron 3 Nano Omni· Nvidia$0.060/1M · 9 August 2026
- Nova 2 Lite (Bedrock)· AWS Bedrock$0.300/1M · 9 August 2026
- Nova Pro 1.0 (Bedrock)· AWS Bedrock$0.799/1M · 9 August 2026
- nova-lite-v1 (Bedrock)· AWS Bedrock$0.059/1M · 9 August 2026
- nova-micro-v1 (Bedrock)· AWS Bedrock$0.035/1M · 9 August 2026
- NVIDIA Nemotron 3 Ultra· Nvidia$0.600/1M · 9 August 2026
- nvidia-nemotron-3-nano-30b-a3b· Nvidia$0.060/1M · 9 August 2026
- o3 Mini High· OpenAI$0.990/1M · 9 August 2026
- o4 Mini High· OpenAI$1.10/1M · 9 August 2026
- Qwen 2.5 72B Instruct· Alibaba$0.357/1M · 9 August 2026
- Qwen2.5 Coder 32B Instruct· Alibaba$0.201/1M · 9 August 2026
- Qwen3 235B A22B FP8· Alibaba$0.200/1M · 9 August 2026
- Qwen3 30B A3B· Alibaba$0.080/1M · 9 August 2026
- Qwen3 30B A3b fp8· Alibaba$0.051/1M · 9 August 2026
- Qwen3 Coder· Alibaba$0.130/1M · 9 August 2026
- Qwen3 Max Thinking· Alibaba$0.780/1M · 9 August 2026
- Qwen3 VL 235B A22B Instruct· Alibaba$0.200/1M · 9 August 2026
- Qwen3 VL 235B A22B Thinking· Alibaba$0.450/1M · 9 August 2026
- Qwen3 VL 30B A3B Instruct· Alibaba$0.130/1M · 9 August 2026
- Qwen3 VL 30B A3B Thinking· Alibaba$0.200/1M · 9 August 2026
- Qwen3 VL 32B Instruct· Alibaba$0.104/1M · 9 August 2026
- Qwen3 VL 8B Instruct· Alibaba$0.080/1M · 9 August 2026
- qwen3-coder-next· Alibaba$0.108/1M · 9 August 2026
- Qwen3.5 9B· Alibaba$0.040/1M · 9 August 2026
- Qwen3.5 Flash· Alibaba$0.090/1M · 9 August 2026
- Qwen3.7 Flash· Alibaba$0.030/1M · 9 August 2026
- Qwen3.8 Max· Alibaba$2.00/1M · 9 August 2026
- Qwq 32B· Alibaba$0.400/1M · 9 August 2026
- WizardLM-2 8x22B· Microsoft$0.493/1M · 9 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.