// Models · price tracker
API price changes
Which model got cheaper or more expensive, where, and when – from daily re-checks of official pages, inference platforms and OpenRouter.
- price cuts
- 18
- increases
- 14
- new listings
- 53
- announced upcoming
- 4
[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object] daily checks since 2026-10-10Updated 2026-10-22
01Announced, not yet in effect
From official pricing pages (checked 2026-10-20). Current prices stay in the table until the effective date.
| What | Effective | Now | From then |
|---|---|---|---|
| DeepSeek Search / Fetch APIBilling starts 2026-11-01 00:00 Beijing for server-side Search/Fetch tool calls when DeepSeek API is used inside DeepSeek Harness, Claude Code or Codex. Token prices of deepseek-flash / deepseek-v4-pro unchanged. | 2026-11-01 | Server-side calls in Harness / Claude Code / Codex not billed yet | Search ¥5 / 1K req · Fetch ¥5 / 1K req |
| Inception Mercury 2 / Mercury 2.5Per inceptionlabs.ai/models: Mercury 2 and 2.5 are no longer available to new users; customers with usage before 2026-10-19 keep access until 2026-11-15; enterprise agreements unaffected. Mercury 3’s launch discount has no published end date. | 2026-11-15 | Closed to new users; existing users (usage before Oct 19) still have access | Retired from Nov 15 → migrate to Mercury 3 ($0.10/$0.35 at 50% launch discount; regular $0.20/$0.70) |
| Gemini 4 ArgonPaid tier only. Batch/Flex remain −50% of the then-current standard price. | 2026-12-01 | $2 / $10 | $4 / $20 |
| Gemini 3.8 FlashPromo through 2026-12-31; free tier remains for many Flash features. | 2027-01-01 | $0.75 / $3.75 | $1.5 / $7.5 |
02Recorded changes, newest first
Prices per 1M tokens in the source currency. Green = cheaper, amber = more expensive.
2026-10-22
| Delisted | minimax/minimax-m2.5 | friendli | |
| Price change | Muse Spark OpenMeta | Nebius | Cache read — → 0.15 USD |
| Price change | Beam 1.0 501B A23BReflection | BaseTen | Cache read 0.13 → 0.04 -69.2%Output 4.3 → 1.2 -72.1%Input 1.3 → 0.4 -69.2% USD |
| Price change | Beam 1.0 501B A23BReflection | Nebius | Cache read — → 0.05 USD |
| Price change | DeepSeek V4.1 Flash 1017DeepSeek | morph | Cache read 0.016 → 0.021 +31.2%Input 0.188 → 0.24 +27.7% USD |
| Price change | GLM 5.3 Flash (batch)Zhipu (Z.ai) · Batch | DeepInfra | Cache read 0.012 → 0.024 +100%Output 0.2 → 0.4 +100%Input 0.06 → 0.12 +100% USD |
| Price change | GLM 5.2Zhipu (Z.ai) | Nebius | Cache read — → 0.14 USD |
| Price change | Kimi K2.7 CodeMoonshot (Kimi) | Nebius | Cache read — → 0.15 USD |
2026-10-21
| Delisted | deepseek/deepseek-v4.1-flash | crusoe | |
| Price change | DeepSeek V4.1 Flash 1017DeepSeek | morph | Cache read 0.007 → 0.016 +128.6%Input 0.055 → 0.188 +241.8% USD |
| Price change | DeepSeek V4.1 Flash 1017DeepSeek | relace | Cache read 0.01 → 0.008 -20%Output 0.024 → 0.5 +1983.3%Input 0.05 → 0.07 +40% USD |
| Price change | GLM 5.3Zhipu (Z.ai) | Novita | Cache read 0.13 → 0.117 -10%Output 2.2 → 1.98 -10%Input 0.7 → 0.63 -10% USD |
| Price change | GLM 5.3 FlashZhipu (Z.ai) | DeepInfra | Cache read 0.015 → 0.03 +100%Output 0.25 → 0.5 +100%Input 0.075 → 0.15 +100% USD |
2026-10-20
| New listing | Mercury 3Inception | official | |
| New listing | MiniMax-M3.1-FlashMiniMax | official | |
| New listing | Mercury 3Inception | inception | |
| Delisted | meta-llama/llama-4-scout | novita | |
| Delisted | meta-llama/llama-4-maverick | novita | |
| Delisted | deepseek/deepseek-r1-0528 | novita | |
| Price change | Nemotron 3.5 Super VLNVIDIA | DeepInfra | Cache read 0.08 → 0.02 -75%Output 0.4 → 0.5 +25%Input 0.1 → 0.15 +50% USD |
| Price change | Beam 1.0 501B A23BReflection | coreweave | Cache read 0.1 → 0.04 -60%Output 3 → 1 -66.7%Input 1 → 0.3 -70% USD |
| Price change | DeepSeek V4.1 Flash 1017DeepSeek | wafer | Output 1.2 → 0.5 -58.3%Input 0.3 → 0.16 -46.7% USD |
| Price change | DeepSeek V4.1 Flash 1017DeepSeek | morph | Output 0.55 → 0.6 +9.1%Input 0.06 → 0.055 -8.3% USD |
| Price change | DeepSeek V4.1 Flash 1017DeepSeek | relace | Cache read 0.008 → 0.01 +25%Output 0.6 → 0.024 -96% USD |
| Price change | DeepSeek V4.1 FlashDeepSeek | Alibaba Cloud | Cache read 0.03 → 0.033 +10%Output 1.2 → 1.32 +10%Input 0.3 → 0.33 +10% USD |
| Price change | Solar Pro 4Upstage | OpenRouter | Cache read 0.018 → 0.03 +66.7%Output 0.36 → 0.6 +66.7%Input 0.09 → 0.15 +66.7% USD |
| Price change | GLM 5.3Zhipu (Z.ai) | Crusoe | Cache read 0.13 → 0.26 +100% USD |
| Price change | GLM 5.3Zhipu (Z.ai) | GMICloud | Cache read 0.182 → 0.13 -28.6%Output 3.08 → 2.2 -28.6%Input 0.98 → 0.7 -28.6% USD |
| Price change | GLM 5.3 FlashZhipu (Z.ai) | GMICloud | Cache read 0.018 → 0.015 -16.7%Output 0.3 → 0.25 -16.7%Input 0.09 → 0.075 -16.7% USD |
| Price change | MiMo-V2.6-ProXiaomi (MiMo) | DeepInfra | Cache read 0.0036 → 0.00324 -10%Output 0.87 → 0.783 -10%Input 0.43 → 0.387 -10% USD |
| Price change | Qwen3.6 35B A3BAlibaba (Qwen) | Parasail | Output 1 → 0.95 -5%Input 0.15 → 0.1 -33.3% USD |
2026-10-19
| New listing | Moondream 3 PreviewMoondream | primeintellect | |
| New listing | Muse Spark OpenMeta | BaseTen | |
| New listing | Muse Spark OpenMeta | Fireworks | |
| New listing | Muse Spark OpenMeta | Nebius | |
| New listing | Muse Spark OpenMeta | Together | |
| New listing | Muse Spark OpenMeta | modal | |
| New listing | Nemotron 3.5 Super VLNVIDIA | venice | |
| New listing | Nemotron 3.5 Super VLNVIDIA | coreweave | |
| New listing | Nemotron 3.5 Super VLNVIDIA | Together | |
| New listing | Nemotron 3.5 Super VLNVIDIA | DeepInfra | |
| New listing | Nemotron 3.5 Super VLNVIDIA | Crusoe | |
| New listing | Beam 1.0 501B A23BReflection | BaseTen | |
| New listing | Beam 1.0 501B A23BReflection | coreweave | |
| New listing | Beam 1.0 501B A23BReflection | Together | |
| New listing | Beam 1.0 501B A23BReflection | Nebius | |
| New listing | DeepSeek V4.1 Flash 1017DeepSeek | DeepSeek (via OpenRouter) | |
| New listing | DeepSeek V4.1 Flash 1017DeepSeek | Parasail | |
| New listing | DeepSeek V4.1 Flash 1017DeepSeek | BaseTen | |
| New listing | DeepSeek V4.1 Flash 1017DeepSeek | primeintellect | |
| New listing | DeepSeek V4.1 Flash 1017DeepSeek | Fireworks | |
| New listing | DeepSeek V4.1 Flash 1017DeepSeek | wafer | |
| New listing | DeepSeek V4.1 Flash 1017DeepSeek | Crusoe | |
| New listing | DeepSeek V4.1 Flash 1017DeepSeek | Cloudflare | |
| New listing | DeepSeek V4.1 Flash 1017DeepSeek | DeepInfra | |
| New listing | DeepSeek V4.1 Flash 1017DeepSeek | inferencenet | |
| New listing | DeepSeek V4.1 Flash 1017DeepSeek | morph | |
| New listing | DeepSeek V4.1 Flash 1017DeepSeek | relace | |
| Price change | Step-5-PreviewStepFun (阶跃星辰) | official | Cache read 0.35 → 0.7 +100% CNY |
| Price change | MiniMax-M2.7-highspeedMiniMax | official | Cache read — → 0.42 CNY |
| Delisted | deepseek/deepseek-v4.1-flash | deepseek | |
| Delisted | deepseek/deepseek-v4.1-flash | cloudflare | |
| Price change | Qwen3.6 35B A3BAlibaba (Qwen) | AtlasCloud | Cache read 0.186 → 0.25 +34.4%Output 1.11375 → 1.25 +12.2%Input 0.186 → 0.25 +34.4% USD |
| Price change | Kimi K3Moonshot (Kimi) | Parasail | Cache read 0.26 → 0.3 +15.4%Output 13 → 15 +15.4%Input 2.6 → 3 +15.4% USD |
| Price change | DeepSeek V4.1 FlashDeepSeek | AtlasCloud | Cache read 0.00384 → 0.0135 +251.6%Output 0.768 → 0.54 -29.7%Input 0.192 → 0.135 -29.7% USD |
2026-10-16
| Price change | GLM 5.3Zhipu (Z.ai) | Crusoe | Cache read 0.26 → 0.13 -50% USD |
2026-10-15
| New listing | SummarizeScaleDown | OpenRouter | |
| New listing | ExtractScaleDown | OpenRouter | |
| Price change | deepseek/deepseek-v4.1-flash | cloudflare | Cache read 0.006 → 0.03 +400%Output 1 → 0.9 -10%Input 0.2 → 0.1 -50% USD |
| Price change | MiMo-V2.6-FlashXiaomi (MiMo) | DeepInfra | Cache read 0.0028 → 0.00266 -5%Output 0.28 → 0.266 -5%Input 0.14 → 0.133 -5% USD |
| Price change | DeepSeek V4.1 FlashDeepSeek | AtlasCloud | Cache read 0.03 → 0.00384 -87.2%Output 1.2 → 0.768 -36%Input 0.3 → 0.192 -36% USD |
2026-10-14
| New listing | Kimi K3Moonshot (Kimi) | Alibaba Cloud | |
| New listing | Kimi K3Moonshot (Kimi) | Moonshot AI (via OpenRouter) | |
| New listing | Kimi K3Moonshot (Kimi) | Amazon Bedrock | |
| New listing | Hy3Tencent (Hunyuan) | Tencent (via OpenRouter) | |
| New listing | DeepSeek V4 Flash 0423DeepSeek | Cloudflare | |
| New listing | DeepSeek V4 Flash 0423DeepSeek | Azure | |
| New listing | DeepSeek V4 Flash 0423DeepSeek | Alibaba Cloud | |
| New listing | GLM 5.3Zhipu (Z.ai) | Z.ai (via OpenRouter) | |
| New listing | GLM 5.3Zhipu (Z.ai) | Cloudflare | |
| New listing | GLM 5.3Zhipu (Z.ai) | Alibaba Cloud | |
| New listing | DeepSeek V4 Flash 0731DeepSeek | Cloudflare | |
| New listing | DeepSeek V4 Flash 0731DeepSeek | Alibaba Cloud | |
| New listing | Hy4 previewTencent (Hunyuan) | Tencent (via OpenRouter) | |
| New listing | MiMo-V2.6-FlashXiaomi (MiMo) | Xiaomi (via OpenRouter) | |
| New listing | GLM 5.3 FlashZhipu (Z.ai) | Z.ai (via OpenRouter) | |
| New listing | deepseek/deepseek-v4.1-flash | deepseek | |
| New listing | DeepSeek V4.1 FlashDeepSeek | Alibaba Cloud | |
| New listing | deepseek/deepseek-v4.1-flash | cloudflare | |
| New listing | Muse WonderMeta | OpenRouter | |
| New listing | Muse Spark 1.3 ContributorMeta | OpenRouter | |
| New listing | Gemini 4 ArgonGoogle | official | |
| Delisted | inclusionai/ling-3.0-flash-vl | novita | |
| Price change | GLM 5.3Zhipu (Z.ai) | SiliconFlow | Cache read 0.208 → 0.13 -37.5%Output 3.52 → 2.2 -37.5%Input 1.12 → 0.7 -37.5% USD |
| Price change | GLM 5.2Zhipu (Z.ai) | SiliconFlow | Cache read 0.221 → 0.13 -41.2%Output 3.74 → 2.2 -41.2%Input 1.19 → 0.7 -41.2% USD |
2026-10-13
| Delisted | qwen/qwen3.8-2.4t-a95b | deepinfra | |
| Price change | Qwen3.8 27BAlibaba (Qwen) | Chutes | Cache read 0.024 → 0.0039 -83.8%Output 2.2 → 1.25 -43.2%Input 0.24 → 0.039 -83.8% USD |
| Price change | Qwen3.6 27BAlibaba (Qwen) | Chutes | Cache read 0.03 → 0.025 -16.7%Input 0.3 → 0.25 -16.7% USD |
| Price change | Kimi K3Moonshot (Kimi) | Chutes | Cache read 0.3 → 0.185 -38.3%Output 15 → 9.5 -36.7%Input 3 → 1.85 -38.3% USD |
| Price change | Kimi K2.6Moonshot (Kimi) | Chutes | Cache read 0.05 → 0.045 -10%Output 2.85 → 1.95 -31.6%Input 0.5 → 0.45 -10% USD |
2026-10-12
| Delisted | z-ai/glm-5.3 | nebius |
FAQ
How do you detect price changes?
Every day we re-read the OpenRouter models and endpoints APIs, re-fetch vendor and platform pricing pages, and diff the result against the previous snapshot. Vendor pages that cannot be parsed automatically are flagged for a manual re-check.
Why do some changes only affect one platform?
Open-weight models are sold by many inference platforms at their own prices; a platform can cut or raise its price without the model vendor changing anything.