This page covers LLM API listed prices in our data for 18-25 October 2026. We tracked 579 models, checked prices 4 times a day, every 6 hours, and logged moves of 5% or more.
Price cuts
12 models had a listed price cut in this period. Qwen3 235B A22B from Cheaper Inference fell to $0.09 input on 24 October 2026, a change of -85%.
| Model | Seller | Input, per 1M | Output, per 1M | Date |
|---|---|---|---|---|
| Qwen3 235B A22B | Cheaper Inference | $0.585 → $0.09-85% | $0.585 → $0.55-6% | 24 October 2026 |
| Mistral Nemo | Cheaper Inference | $0.0975 → $0.029-70% | $0.0975 → $0.03-69% | 24 October 2026 |
| Qwen3.8 27B | Ionstream | $0.0893 → $0.04-55% | $2.35 → $1.95-17% | 19 October 2026 |
| Qwen3.5-9B | PrimeIntellect | $0.18 → $0.17-6% | $0.54 → $0.25-54% | 25 October 2026 |
| GLM 5.2 | Decart | $2.50 → $1.20-52% | $9.00 → $8.00-11% | 20 October 2026 |
| Gemma 3 4B | OpenRouter | $0.05 → $0.04-20% | $0.10 → $0.05-50% | 20 October 2026 |
| Ling 3.1 Flash | DeepInfra | $0.15 → $0.075-50% | $0.45 → $0.225-50% | 20 October 2026 |
| DeepSeek V4 Pro 0423 | Relace | $0.90 → $0.90unchanged | $3.00 → $1.50-50% | 19 October 2026 |
| Qwen3.5-27B | OpenRouter | $0.26 → $0.195-25% | $2.60 → $1.56-40% | 19 October 2026 |
| Qwen3.6 35B A3B | OpenRouter | $0.15 → $0.10-33% | $1.00 → $0.95-5% | 20 October 2026 |
| GLM 5.3 | PrimeIntellect | $1.40 → $0.98-30% | $4.40 → $3.08-30% | 24 October 2026 |
| GLM 5.3 | Fireworks | $1.40 → $0.98-30% | $4.40 → $3.08-30% | 22 October 2026 |
Price rises
12 models had a listed price rise in this period. GLM 5.3 Flash from DeepInfra went to $0.15 input and $0.50 output on 21 October 2026, a change of +100%.
| Model | Seller | Input, per 1M | Output, per 1M | Date |
|---|---|---|---|---|
| GLM 5.3 Flash | DeepInfra | $0.075 → $0.15+100% | $0.25 → $0.50+100% | 21 October 2026 |
| Mistral Large 4 | OpenRouter | $0.68 → $1.36+100% | $2.09 → $4.18+100% | 20 October 2026 |
| Solar Pro 4 | OpenRouter | $0.09 → $0.15+67% | $0.36 → $0.60+67% | 20 October 2026 |
| Kimi K3 | Makora | $1.53 → $0.65-58% | $12.75 → $15.00+18% | 20 October 2026 |
| Mistral Codestral | Cheaper Inference | $0.195 → $0.30+54% | $0.585 → $0.90+54% | 24 October 2026 |
| mistral-large-4 | Cheaper Inference | $0.442 → $0.68+54% | $1.36 → $2.09+54% | 24 October 2026 |
| Mistral Medium 3.5 | Cheaper Inference | $0.975 → $1.50+54% | $4.88 → $7.50+54% | 24 October 2026 |
| meta/meta/muse-wonder-1.0 | Cheaper Inference | $2.60 → $4.00+54% | $13.00 → $20.00+54% | 24 October 2026 |
| claude-opus-5.5-fast | Cheaper Inference | $5.20 → $8.00+54% | $26.00 → $40.00+54% | 24 October 2026 |
| gpt-5.1-thinking | Cheaper Inference | $0.8125 → $1.25+54% | $6.50 → $10.00+54% | 24 October 2026 |
| o1 | Cheaper Inference | $9.75 → $15.00+54% | $39.00 → $60.00+54% | 24 October 2026 |
| DeepSeek V4.1 Flash (Fast) | Cheaper Inference | $0.39 → $0.60+54% | $1.56 → $2.40+54% | 24 October 2026 |
New models
OpenRouter added 2 models this week. MiniMax M3.1 Flash arrived on 19 October 2026 and Kimi K3.1 Pro on 23 October 2026.
| Model | Added to OpenRouter | Sellers now | Cheapest input | Cheapest output |
|---|---|---|---|---|
| MiniMax M3.1 Flash | 19 October 2026 | 20 | $0.1125 | $0.45 |
| Kimi K3.1 Pro | 23 October 2026 | 17 | $2.00 | $10.00 |
Cheapest right now
This table shows the cheapest seller for each model right now, with listed input and output prices and the number of sellers.
| Model | Cheapest seller | Input, per 1M | Output, per 1M | Sellers |
|---|---|---|---|---|
| DeepSeek V4.1 Flash#1 on OpenRouter, 124T tokens in 30 days | engy | $0.03 | $0.12 | 36 |
| GLM 5.3 Flash#3 on OpenRouter, 49.7T tokens in 30 days | Relace | $0.0513 | $0.0701 | 77 |
| MiMo-V2.6-Flash#4 on OpenRouter, 40.4T tokens in 30 days | Nous Research | $0.091 | $0.182 | 12 |
| Hy4 preview#5 on OpenRouter, 33T tokens in 30 days | Nous Research | $0.417 | $1.25 | 16 |
| GPT-6 Luna#7 on OpenRouter, 28.2T tokens in 30 days | Cheaper Inference | $0.065 | $0.325 | 23 |
| Step 5 Preview#9 on OpenRouter, 23.1T tokens in 30 days | OpenRouter | $1.00 | $2.70 | 13 |
Prices move, so check the model pages and the price changes page before you commit spend. All price changes.
How we count: standard list prices in US dollars per million tokens, as each seller publishes them, checked every 6 hours. A change is listed when it held for at least two checks and moved input or output by 5% or more; sellers we started tracking during the week are left out. The text was written by DeepSeek V4.1 Flash from these figures, and every number in it was checked against our data before publishing. The tables are generated straight from our data.