AI API & GPU price report: June 2026
In June 2026 we recorded 47 price changes across AI model APIs and GPU clouds: 25 API price drops, 22 increases and 0 GPU rental price changes. The biggest cut was Mistral Large 3 on Mistral AI (−75%).
Price changes
47
API drops
25
API increases
22
H100 lowest
$1.47/hr
GPU price index
On-demand price per GPU-hour across the GPU clouds we track, at the end of June 2026, compared with the start of the month.
| GPU | Lowest | Cheapest on | Median | Median vs start | Offers |
|---|---|---|---|---|---|
| B200 | $1.99 | Vultr | $6.49 | no change | 6 |
| H200 | $2.99 | Verda | $3.99 | no change | 7 |
| H100 | $1.47 | Vast.ai | $2.29 | no change | 19 |
| A100 80GB | $1.19 | RunPod | $1.57 | no change | 13 |
| L40S | $0.79 | RunPod | $1.00 | no change | 5 |
| RTX PRO 6000 | $1.71 | Latitude.sh | $1.80 | no change | 3 |
| RTX 5090 | $0.30 | Vast.ai | $0.30 | no change | 1 |
| RTX 4090 | $0.24 | Vast.ai | $0.34 | no change | 4 |
| MI300X | $1.85 | Vultr | $2.65 | no change | 2 |
Biggest API price drops
Price drops(25)
- Mistral Large 3onMistral AIinput $2 → $0.5/1M, output $6 → $1.5/1M↓−75%
- Mistral Large 3onAmazon Bedrockinput $8 → $3/1M, output $24 → $9/1M↓−63%
- Qwen 3 Max ThinkingonAlibaba Cloudinput $3 → $1.2/1M, output $15 → $6/1M↓−60%
- Grok 4.20onxAIinput $2 → $1.25/1M, output $6 → $2.5/1M↓−58%
- Grok 4.20onOpenRouterinput $2 → $1.25/1M, output $6 → $2.5/1M↓−58%
- Qwen 3.5 35BonDeepInfrainput $0.22 → $0.14/1M, output $2.2 → $1/1M↓−55%
- MiniMax M2.5onOpenRouterinput $0.118 → $0.12/1M, output $0.99 → $0.48/1M↓−52%
- DeepSeek R1onDeepSeekinput $0.28 → $0.14/1M, output $0.42 → $0.28/1M↓−50%
- DeepSeek V3.2onDeepSeekinput $0.28 → $0.14/1M, output $0.42 → $0.28/1M↓−50%
- Mistral Small 4onParasailoutput $0.6 → $0.3/1M↓−50%
…and 15 more; see the daily pages under price changes.
Biggest API price increases
Price increases(22)
- GPT-5.4 MinionWaveSpeed AIinput $0.4 → $0.75/1M, output $1.6 → $4.5/1M↑+181%
- Qwen 3.5 9BonDeepInfrainput $0.04 → $0.1/1M, output $0.2 → $0.15/1M↑+150%
- Gemini 3.1 ProonWaveSpeed AIinput $1.25 → $2/1M, output $5 → $12/1M↑+140%
- Kimi K2.5onSiliconFlowinput $0.23 → $0.45/1M, output $3 → $2.25/1M↑+96%
- Qwen 3 235BonAlibaba Cloudinput $0.4 → $0.7/1M, output $2.4 → $2.8/1M↑+75%
- Qwen 3 8BonAlibaba Cloudoutput $0.4 → $0.7/1M↑+75%
- Qwen 3.5 9BonTogether AIinput $0.1 → $0.17/1M, output $0.15 → $0.25/1M↑+70%
- Command R+onAmazon Bedrockinput $2.5 → $3/1M, output $10 → $15/1M↑+50%
- Ministral 3 8BonMistral AIinput $0.1 → $0.15/1M, output $0.1 → $0.15/1M↑+50%
- Qwen 3 235BonParasailinput $0.1 → $0.14/1M, output $0.6 → $0.8/1M↑+40%
Most active providers
- OpenRouter14 changes
- DeepInfra6 changes
- Alibaba Cloud4 changes
- WaveSpeed AI4 changes
- Mistral AI3 changes
- Amazon Bedrock2 changes
- DeepSeek2 changes
- fal.ai2 changes
- Novita AI2 changes
- Parasail2 changes
Using these numbers
You're welcome to quote this report. Please link to it as the source:
Source: Inference Hub, AI API & GPU Price Report, June 2026 (https://inferencehub.org/price-report/2026-06)
How we collect prices: from each provider's own pricing page or API, checked weekly; see our methodology. Percentages compare the new price with the old; when input and output prices both moved, the larger move is shown. Get daily alerts by RSS or on Telegram.