LLM hosting prices: what changed in October 2026
196 price cuts and 99 rises; the largest is 100% on DeepInfra zai-org/GLM-5.3-Flash. 300 models newly sold per token, 225 models withdrawn, 124 retirements this month and 45 new GPU instances.
196 price cuts · 99 price rises · 300 models newly sold per token · 225 withdrawn · 45 new GPU instances · 126 new regions
Biggest price moves
| Cloud | What | Price | Change |
|---|---|---|---|
| DeepInfra | zai-org/GLM-5.3-Flash, batch, global | $0.012 → $0.024 | +100% |
| DeepInfra | zai-org/GLM-5.3-Flash, standard, global | $0.015 → $0.03 | +100% |
| DeepInfra | zai-org/GLM-5.3-Flash, batch, global | $0.20 → $0.40 per 1M output tokens | +100% |
| DeepInfra | zai-org/GLM-5.3-Flash, standard, global | $0.25 → $0.50 per 1M output tokens | +100% |
| DeepInfra | zai-org/GLM-5.3-Flash, batch, global | $0.06 → $0.12 per 1M input tokens | +100% |
| DeepInfra | zai-org/GLM-5.3-Flash, standard, global | $0.075 → $0.15 per 1M input tokens | +100% |
| Mistral | mistral-large-4, global | $2.09 → $4.18 per 1M output tokens | +100% |
| Mistral | mistral-large-4, global | $0.07 → $0.14 | +100% |
| Mistral | mistral-large-4, global | $0.68 → $1.36 per 1M input tokens | +100% |
| DeepInfra | nvidia/NVIDIA-Nemotron-3.5-Super-VL, batch, global | $0.064 → $0.016 | -75% |
And 285 more.
Models newly sold per token
- OpenAI: gpt-rosalind-research from $5.00 in / $25.00 out per 1M tokens
- OpenAI: gpt-rosalind-discovery from $5.00 in / $25.00 out per 1M tokens
- OpenAI: gpt-6-rosalind from $15.00 in / $75.00 out per 1M tokens
- Snowflake: openai-gpt-6.1-terra from $1.00 in / $5.00 out per 1M tokens
- Azure: 61 terra ShortCo Opt std Gl
- Azure: 61 terra ShortCo Opt Std DZ
- Azure: 61 terra ShortCo Inp Std Gl from $1.00 in / out per 1M tokens
- Azure: 61 terra ShortCo Inp Std DZ from $1.10 in / out per 1M tokens
- Azure: 61 terra ShortCo Cd Wr Std Gl
- Azure: 61 terra ShortCo Cd Wr Std DZ
- Azure: 61 terra ShortCo Cd Inp Std Gl
- Azure: 61 terra ShortCo Cd Inp Std DZ
- Azure: 61 terra LongCo Opt std Gl
- Azure: 61 terra LongCo Opt Std DZ
- Azure: 61 terra LongCo Inp Std Gl from $2.00 in / out per 1M tokens
- Azure: 61 terra LongCo Inp Std DZ from $2.20 in / out per 1M tokens
- Azure: 61 terra LongCo Cd Wr Std Gl
- Azure: 61 terra LongCo Cd Wr Std DZ
- Azure: 61 terra LongCo Cd Inp Std Gl
- Azure: 61 terra LongCo Cd Inp Std DZ
- Fireworks AI: Ultramarine 1 Flash from $0.1125 in / $0.375 out per 1M tokens
- Mistral: deepseek-v4-1-flash-1017 from $0.30 in / $1.20 out per 1M tokens
- Databricks: GPT-6.1 Terra from $1.00 in / $5.00 out per 1M tokens
- Fireworks AI: Rune 26B-A4B from $0.021 in / $0.00 out per 1M tokens
- Huawei Cloud: DeepSeek-V4.1-Flash-1017 (Text chat and image understanding are supported.) from $0.15 in / $0.60 out per 1M tokens
- Huawei Cloud: DeepSeek-V4.1-Flash (Text chat and image understanding are supported.) from $0.15 in / $0.60 out per 1M tokens
- Google Cloud: Gemini ER 2.0 from $1.00 in / $5.00 out per 1M tokens
- Nebius: zai-org/GLM-5.2 from $1.40 in / $4.40 out per 1M tokens
- Alibaba Cloud: qwen3.8-omni-flash from $0.113 in / $0.382 out per 1M tokens
- OpenAI: gpt-6.1-terra from $1.00 in / $5.00 out per 1M tokens
- Crusoe: Qwen3.8 27B text, image from $0.25 in / $2.00 out per 1M tokens
- Crusoe: Nemotron 3.5 Super VL 120B A12B text, image, video from $0.10 in / $0.40 out per 1M tokens
- Crusoe: Nemotron 3.5 Lightning 30B A3B from $0.05 in / $0.20 out per 1M tokens
- Crusoe: Nemotron 3 Super 120B from $0.30 in / $2.40 out per 1M tokens
- Crusoe: Nemotron 3 Nano Omni text, image from $0.30 in / $1.83 out per 1M tokens
- Crusoe: Nemotron 3 Nano 30B from $0.05 in / $0.20 out per 1M tokens
- Crusoe: Gemma 4 31B text, image from $0.14 in / $0.40 out per 1M tokens
- Crusoe: GLM-5.3 Flash text, image from $0.15 in / $0.50 out per 1M tokens
- Crusoe: GLM-5.3 from $1.40 in / $4.40 out per 1M tokens
- Crusoe: DeepSeek V4.1 Flash text, image from $0.29 in / $1.20 out per 1M tokens
- Crusoe: DeepSeek V4 Flash 0731 from $0.14 in / $0.28 out per 1M tokens
- Nebius: moonshotai/Kimi-K3 from $3.00 in / $15.00 out per 1M tokens
- Nebius: deepseek-ai/DeepSeek-V4.1-Flash-1017 from $0.30 in / $1.20 out per 1M tokens
- Nebius: MiniMaxAI/MiniMax-M3.1-Flash from $0.15 in / $0.60 out per 1M tokens
- Huawei Cloud: DeepSeek-V4.1-Flash (Text and image are supported) from $0.15 in / $0.60 out per 1M tokens
- Databricks: Gemini 4 Argon from $2.50 in / $12.50 out per 1M tokens
- Databricks: DeepSeek V4.1 Flash (1017) from $0.30 in / $1.20 out per 1M tokens
- Databricks: DeepSeek V4.1 Flash (0910) from $0.30 in / $1.20 out per 1M tokens
- Together AI: MiniMaxAI/MiniMax-M3.1-Flash from $0.15 in / $0.60 out per 1M tokens
- Mistral: zai-glm-5-3-flash from $0.15 in / $0.50 out per 1M tokens
- Mistral: voxtral-small-2507 from $0.10 in / $0.40 out per 1M tokens
- Azure: gpt-rt-21 Text opt Rg
- Azure: gpt-rt-21 Text inp Rg from $5.20 in / out per 1M tokens
- Azure: gpt-rt-21 Text cd inp Rg
- Azure: gpt-rt-21 Image inp Rg from $6.50 in / out per 1M tokens
- Azure: gpt-rt-21 Image cd inp Rg
- Azure: gpt-rt-21 Audio opt Rg
- Azure: gpt-rt-21 Audio inp Rg from $41.60 in / out per 1M tokens
- Azure: gpt-rt-21 Audio cd inp Rg
- Azure: chat-latest 10072026 opt Gl
- Azure: chat-latest 10072026 opt Dz
- Azure: chat-latest 10072026 inp Gl from $5.00 in / out per 1M tokens
- Azure: chat-latest 10072026 inp Dz from $5.50 in / out per 1M tokens
- Azure: chat-latest 10072026 cd inp Gl
- Azure: chat-latest 10072026 cd inp Dz
- Azure: MAI code flash 11 Opt FT Gl
- Azure: MAI code flash 11 Inp FT Gl from $0.20 in / out per 1M tokens
- Azure: MAI code flash 11 Cd FT Gl
- Azure: Fal Test MM H3 Mx Txt-Vd 768
- Azure: Fal Test MM H3 Mx Txt-Vd 480
- Azure: Fal Test MM H3 Mx Txt-Vd 1080
- Azure: Fal Test MM H3 Mx Trb Txt-Vd 768
- Azure: Fal Test MM H3 Mx Trb Txt-Vd 480
- Azure: Fal Test MM H3 Mx Trb Txt-Vd 1080
- Azure: Fal Test MM H3 Mx Trb Im-Vd 768
- Azure: Fal Test MM H3 Mx Trb Im-Vd 480
- Azure: Fal Test MM H3 Mx Trb Im-Vd 1080
- Azure: Fal Test MM H3 Mx Ref-Vd 768
- Azure: Fal Test MM H3 Mx Ref-Vd 480
- Azure: Fal Test MM H3 Mx Ref-Vd 1080
- Azure: Fal Test MM H3 Mx Im-Vd 768
- Azure: Fal Test MM H3 Mx Im-Vd 480
- Azure: Fal Test MM H3 Mx Im-Vd 1080
- Azure: Fal Test LTX 2.5 Pro Txt-Vd 720
- Azure: Fal Test LTX 2.5 Pro Im-Vd 720
- Azure: Fal Test LTX 2.5 Pro Im-Vd 1080
- Azure: Fal Test LTX 2.5 Fast Txt-Vd 720
- Azure: Fal Test LTX 2.5 Fast Txt-Vd 4K
- Azure: Fal Test LTX 2.5 Fast Txt-Vd 1440
- Azure: Fal Test LTX 2.5 Fast Txt-Vd 1080
- Azure: Fal Test LTX 2.5 Fast Im-Vd 4K
- Azure: Fal Test LTX 2.5 Fast Im-Vd 1440
- Azure: Fal Test LTX 2.5 Fast Im-Vd 1080
- Alibaba Cloud: glm-5.3-prime from $2.80 in / $8.80 out per 1M tokens
- Crusoe: Qwen3 235B A22B Instruct 2507 from $0.22 in / $0.80 out per 1M tokens
- Fireworks AI: Muse Spark Open from $0.625 in / $2.12 out per 1M tokens
- Nebius: moonshotai/Kimi-K3 from $3.00 in / $15.00 out per 1M tokens
- Cloudflare: deepseek-ai/deepseek-v4.1-flash-1017 from $0.30 in / $1.20 out per 1M tokens
- Together AI: deepseek-ai/DeepSeek-V4.1-Flash-1017 from $0.30 in / $1.20 out per 1M tokens
- Databricks: GPT-5.6 Luna from $0.40 in / $2.40 out per 1M tokens
- Alibaba Cloud: deepseek-v4.1-flash-1017 from $0.283 in / $1.13 out per 1M tokens
- Azure: Qwen38 27b SFT Gl
- Azure: Qwen38 27b SFT Dz
- Azure: Qwen36 35b SFT Gl
- Azure: Qwen36 35b SFT Dz
- Azure: Muse Glimmer SFT Gl
- Azure: Muse Glimmer SFT Dz
- Azure: Image 2.6 Text Input DZ from $5.50 in / out per 1M tokens
- Azure: Image 2.6 Image Output DZ
- Azure: Image 2.6 Image Input DZ from $8.80 in / out per 1M tokens
- Azure: Image 2.6 Flash Text In DZ from $1.93 in / out per 1M tokens
- Azure: Image 2.6 Flash Image Output DZ
- Azure: Image 2.6 Flash Image Input DZ from $2.75 in / out per 1M tokens
- Azure: Image 2.5 Pro Text Input DZ from $5.50 in / out per 1M tokens
- Azure: Image 2.5 Pro Image Output DZ
- Azure: Image 2.5 Pro Image Input DZ from $8.80 in / out per 1M tokens
- Azure: GPT OSS 120b SFT Gl
- Azure: GPT OSS 120b SFT Dz
- Azure: Fal MM H3 Max Txt to Vd 768p
- Azure: Fal MM H3 Max Txt to Vd 480p
- Azure: Fal MM H3 Max Txt to Vd 1080p
- Azure: Fal MM H3 Max Turbo Txt to Vd 768p
- Azure: Fal MM H3 Max Turbo Txt to Vd 480p
- Azure: Fal MM H3 Max Turbo Txt to Vd 1080p
- Azure: Fal MM H3 Max Turbo Img to Vd 768p
- Azure: Fal MM H3 Max Turbo Img to Vd 480p
- Azure: Fal MM H3 Max Turbo Img to Vd 1080p
- Azure: Fal MM H3 Max Ref to Vd 768p
- Azure: Fal MM H3 Max Ref to Vd 480p
- Azure: Fal MM H3 Max Ref to Vd 1080p
- Azure: Fal MM H3 Max Img to Vd 768p
- Azure: Fal MM H3 Max Img to Vd 480p
- Azure: Fal MM H3 Max Img to Vd 1080p
- Azure: Fal LTX 2.5 Pro Txt to Vd 720p
- Azure: Fal LTX 2.5 Pro Img to Vd 720p
- Azure: Fal LTX 2.5 Pro Img to Vd 1080p
- Azure: Fal LTX 2.5 Fast Txt to Vd 720p
- Azure: Fal LTX 2.5 Fast Txt to Vd 4K
- Azure: Fal LTX 2.5 Fast Txt to Vd 1440p
- Azure: Fal LTX 2.5 Fast Txt to Vd 1080p
- Azure: Fal LTX 2.5 Fast Img to Vd 4K
- Azure: Fal LTX 2.5 Fast Img to Vd 1440p
- Azure: Fal LTX 2.5 Fast Img to Vd 1080p
- Azure: FW DS V4.1 Flash Opt - US DZ
- Azure: FW DS V4.1 Flash Inp - US DZ from $0.45 in / out per 1M tokens
- Azure: FW DS V4.1 Flash Cd Inp - US DZ
- Azure: Cyber 1.1 Opt glbl
- Azure: Cyber 1.1 Inp glbl from $0.60 in / out per 1M tokens
- Azure: Cyber 1.1 Cd Inp glbl
- Fireworks AI: DeepSeek V4.1 Flash 1017 from $0.15 in / $0.60 out per 1M tokens
- Together AI: reflection/Beam-1.0-501B-A23B from $0.40 in / $1.20 out per 1M tokens
- Nebius: reflection/beam from $0.25 in / $1.00 out per 1M tokens
- Cloudflare: daily-co/phonellm-mini-alpha-2 from $0.06 in / $0.24 out per 1M tokens
- Together AI: meta-models/Muse-Spark-Open-816B-A42B from $1.25 in / $4.25 out per 1M tokens
- Nebius: nvidia/Nemotron-3.5-Super-VL from $0.30 in / $0.90 out per 1M tokens
- Nebius: meta-models/muse-spark-open from $1.25 in / $4.25 out per 1M tokens
- Google Cloud: DeepSeek-V4.1-Flash from $0.30 in / $1.20 out per 1M tokens
- Google Cloud: Gemini 4 Argon from $2.00 in / $10.00 out per 1M tokens
- Core42: qwen-3.8-27b from $0.045 in / $2.30 out per 1M tokens
- Azure: 6.1-sol ShortCo Opt Flex Gl
- Azure: 6.1-sol ShortCo Inp Flex Gl from $1.00 in / out per 1M tokens
- Azure: 6.1-sol ShortCo Cd Wr Flex Gl
- Azure: 6.1-sol ShortCo Cd Inp Flex Gl
- Azure: 6.1-sol LongCo Opt Flex Gl
- Azure: 6.1-sol LongCo Inp Flex Gl from $2.00 in / out per 1M tokens
- Azure: 6.1-sol LongCo Cd Wr Flex Gl
- Azure: 6.1-sol LongCo Cd Inp Flex Gl
- Azure: 6-sol ShortCo Opt Flex Gl
- Azure: 6-sol ShortCo Inp Flex Gl from $1.00 in / out per 1M tokens
- Azure: 6-sol ShortCo Cd Wr Flex Gl
- Azure: 6-sol ShortCo Cd Inp Flex Gl
- Azure: 6-sol LongCo Opt Flex Gl
- Azure: 6-sol LongCo Inp Flex Gl from $2.00 in / out per 1M tokens
- Azure: 6-sol LongCo Cd Wr Flex Gl
- Azure: 6-sol LongCo Cd Inp Flex Gl
- Azure: 6-luna ShortCo Opt Flex Gl
- Azure: 6-luna ShortCo Inp Flex Gl from $0.05 in / out per 1M tokens
- Azure: 6-luna ShortCo Cd Wr Flex Gl
- Azure: 6-luna ShortCo Cd Inp Flex Gl
- Azure: 6-luna LongCo Opt Flex Gl
- Azure: 6-luna LongCo Inp Flex Gl from $0.10 in / out per 1M tokens
- Azure: 6-luna LongCo Cd Wr Flex Gl
- Azure: 6-luna LongCo Cd Inp Flex Gl
- Azure: 6-astra ShortCo Opt Flex Gl
- Azure: 6-astra ShortCo Inp Flex Gl from $5.00 in / out per 1M tokens
- Azure: 6-astra ShortCo Cd Wr Flex Gl
- Azure: 6-astra ShortCo Cd Inp Flex Gl
- Azure: 6-astra LongCo Opt Flex Gl
- Azure: 6-astra LongCo Inp Flex Gl from $10.00 in / out per 1M tokens
- Azure: 6-astra LongCo Cd Wr Flex Gl
- Azure: 6-astra LongCo Cd Inp Flex Gl
- Azure: 4.7 Outp glbl L
- Azure: 4.7 Outp glbl
- Azure: 4.7 Outp DZ L
- Azure: 4.7 Outp DZ
- Azure: 4.7 Inp glbl L from $4.00 in / out per 1M tokens
- Azure: 4.7 Inp glbl from $2.00 in / out per 1M tokens
- Azure: 4.7 Inp DZ L from $4.40 in / out per 1M tokens
- Azure: 4.7 Inp DZ from $2.20 in / out per 1M tokens
- Azure: 4.7 Cached glbl L
- Azure: 4.7 Cached glbl
- Azure: 4.7 Cached DZ L
- Azure: 4.7 Cached DZ
- Azure: V4.1 Flash cached Glbl
- Azure: V4.1 Flash cached DZ
- Azure: V4.1 Flash Output Glbl
- Azure: V4.1 Flash Output DZ
- Azure: V4.1 Flash Input Glbl from $0.30 in / out per 1M tokens
- Azure: V4.1 Flash Input DZ from $0.33 in / out per 1M tokens
- Azure: Microsoft Decision 1
- Azure: Embed V5 Pro Txt Glbl
- Azure: Embed V5 Pro Txt DZ
- Azure: Embed V5 Pro Img Glbl
- Azure: Embed V5 Pro Img DZ
- Azure: Embed V5 Fast Txt Glbl
- Azure: Embed V5 Fast Txt DZ
- Azure: Embed V5 Fast Img Glbl
- Azure: Embed V5 Fast Img DZ
- Azure: 6.1-sol ShortCo Opt Std Gl
- Azure: 6.1-sol ShortCo Opt Std DZ
- Azure: 6.1-sol ShortCo Inp Std Gl from $2.00 in / out per 1M tokens
- Azure: 6.1-sol ShortCo Inp Std DZ from $2.20 in / out per 1M tokens
- Azure: 6.1-sol ShortCo Cd Wr Std Gl
- Azure: 6.1-sol ShortCo Cd Wr Std DZ
- Azure: 6.1-sol ShortCo Cd Inp Std Gl
- Azure: 6.1-sol ShortCo Cd Inp Std DZ
- Azure: 6.1-sol LongCo Opt Std Gl
- Azure: 6.1-sol LongCo Opt Std DZ
- Azure: 6.1-sol LongCo Inp Std Gl from $4.00 in / out per 1M tokens
- Azure: 6.1-sol LongCo Inp Std DZ from $4.40 in / out per 1M tokens
- Azure: 6.1-sol LongCo Cd Wr Std Gl
- Azure: 6.1-sol LongCo Cd Wr Std DZ
- Azure: 6.1-sol LongCo Cd Inp Std Gl
- Azure: 6.1-sol LongCo Cd Inp Std DZ
- AWS: TwelveLabs Pegasus 1.5
- OCI: OCI - Generative AI - xAI - Grok 4.3 Priority - Output Tokens less than 200K Tokens
- OCI: OCI - Generative AI - xAI - Grok 4.3 Priority - Output Tokens greater than 200K Tokens
- OCI: OCI - Generative AI - xAI - Grok 4.3 Priority - Input Tokens less than 200K Tokens from $2.50 in / out per 1M tokens
- OCI: OCI - Generative AI - xAI - Grok 4.3 Priority - Input Tokens greater than 200K Tokens from $5.00 in / out per 1M tokens
- OCI: OCI - Generative AI - xAI - Grok 4.3 Priority - Cached Input Tokens less than 200K Tokens
- OCI: OCI - Generative AI - xAI - Grok 4.3 Priority - Cached Input Tokens greater than 200K Tokens
- OCI: OCI - Generative AI - xAI - Grok 4.2 Priority - Output Tokens less than 200K Tokens
- OCI: OCI - Generative AI - xAI - Grok 4.2 Priority - Output Tokens greater than 200K Tokens
- OCI: OCI - Generative AI - xAI - Grok 4.2 Priority - Input Tokens less than 200K Tokens from $2.50 in / out per 1M tokens
- OCI: OCI - Generative AI - xAI - Grok 4.2 Priority - Input Tokens greater than 200K Tokens from $5.00 in / out per 1M tokens
- OCI: OCI - Generative AI - xAI - Grok 4.2 Priority - Cached Input Tokens less than 200K Tokens
- OCI: OCI - Generative AI - xAI - Grok 4.2 Priority - Cached Input Tokens greater than 200K Tokens
- AWS: SI-Ultra-V1
- AWS: SI-Core-V1
- AWS: SD3LargeV1
- AWS: SD3.5LargeV1
- AWS: Rerank 3.5
- AWS: Pegasus 1.2
- AWS: Palmyra X5 from $0.60 in / $6.00 out per 1M tokens
- AWS: Palmyra X4 from $2.50 in / $10.00 out per 1M tokens
- AWS: OpenAI GPT-6 Astra from $10.00 in / $50.00 out per 1M tokens
- AWS: Marengo Embed 3.0
- AWS: Marengo Embed 2.7
- AWS: Llama2-70B from $1.95 in / $2.56 out per 1M tokens
- AWS: Llama2-13B from $0.75 in / $1.00 out per 1M tokens
- AWS: Jurassic Ultra from $18.80 in / $18.80 out per 1M tokens
- AWS: Jurassic Mid from $12.50 in / $12.50 out per 1M tokens
- AWS: Jamba-Instruct from $0.50 in / $0.70 out per 1M tokens
- AWS: Jamba 1.5 Mini from $0.20 in / $0.40 out per 1M tokens
- AWS: Jamba 1.5 Large from $2.00 in / $8.00 out per 1M tokens
- AWS: Embed v4 from $0.12 in / out per 1M tokens
- AWS: Embed Multilingual from $0.10 in / out per 1M tokens
- AWS: Embed English from $0.10 in / out per 1M tokens
- AWS: Command-Light from $0.30 in / $0.60 out per 1M tokens
- AWS: Command R+ from $3.00 in / $15.00 out per 1M tokens
- AWS: Command R from $0.50 in / $1.50 out per 1M tokens
- AWS: Command from $1.00 in / $2.00 out per 1M tokens
- AWS: Claude Sonnet 5.5 from $2.00 in / $10.00 out per 1M tokens
- AWS: Claude Sonnet 5 from $2.00 in / $10.00 out per 1M tokens
- AWS: Claude Sonnet 4.6 from $1.50 in / $7.50 out per 1M tokens
- AWS: Claude Sonnet 4.5 from $1.50 in / $7.50 out per 1M tokens
- AWS: Claude Opus 5.5 from $4.00 in / $20.00 out per 1M tokens
- AWS: Claude Opus 5 from $2.50 in / $12.50 out per 1M tokens
- AWS: Claude Opus 4.8 from $5.00 in / $25.00 out per 1M tokens
- AWS: Claude Opus 4.7 from $5.00 in / $25.00 out per 1M tokens
- AWS: Claude Opus 4.6 from $2.50 in / $12.50 out per 1M tokens
- AWS: Claude Opus 4.5 from $2.50 in / $12.50 out per 1M tokens
- AWS: Claude Opus 4.1 from $15.00 in / $75.00 out per 1M tokens
- AWS: Claude Opus 4 from $7.50 in / $37.50 out per 1M tokens
- AWS: Claude Mythos Preview from $27.50 in / $137.50 out per 1M tokens
- AWS: Claude Mythos 5.1 from $10.00 in / $50.00 out per 1M tokens
- AWS: Claude Mythos 5 from $11.00 in / $55.00 out per 1M tokens
- AWS: Claude Haiku 5.5 from $0.10 in / $0.50 out per 1M tokens
- AWS: Claude Fable 5.1 from $10.00 in / $50.00 out per 1M tokens
- AWS: Claude Fable 5 from $10.00 in / $50.00 out per 1M tokens
- AWS: Claude 3.7 Sonnet from $1.50 in / $7.50 out per 1M tokens
- AWS: Claude 3.5 Sonnet v2 from $1.50 in / $7.50 out per 1M tokens
- AWS: Claude 3.5 Sonnet from $1.50 in / $7.50 out per 1M tokens
- AWS: Claude 3.5 Haiku from $0.40 in / $2.00 out per 1M tokens
- AWS: Claude 3 Opus from $7.50 in / $37.50 out per 1M tokens
- Anthropic: Claude Haiku 5.5 from $0.25 in / $1.25 out per 1M tokens
- AWS: Claude Haiku 5.5 (Amazon Bedrock Edition) from $0.10 in / $0.50 out per 1M tokens
- Google Cloud: Gemini Nano Banana 2.1 from $1.50 in / $7.50 out per 1M tokens
- AWS: Nova 2.5 Sonic from $0.319 in / $2.65 out per 1M tokens
- Core42: gpt-6-sol from $2.20 in / $11.00 out per 1M tokens
Withdrawn and retiring
- Together AI no longer sells Qwen/Qwen3.8-2.4T-A95B
- Mistral no longer sells mistral-medium-3-5
- Mistral no longer sells mistral-large-2512
- Nebius no longer sells zai-org/GLM-5.3-Flash
- Nebius no longer sells reflection/beam
- Nebius no longer sells deepseek-ai/DeepSeek-V4.1-Flash
- DeepInfra no longer sells Gryphe/MythoMax-L2-13b
- DeepInfra no longer sells stepfun-ai/Step-5-Preview
- Google Cloud no longer sells gpt-oss-20b
- Google Cloud no longer sells gpt-oss-120b
- Google Cloud no longer sells Qwen3-Next-80B-Thinking
- Google Cloud no longer sells Qwen3-Next-80B-Instruct
- Google Cloud no longer sells Qwen3-Coder-480B-A35B-Instruct
- Google Cloud no longer sells Qwen3-235B-A22B-Instruct-2507
- Google Cloud no longer sells Mistral Small 3.1 (25.03)
- Google Cloud no longer sells Mistral Medium 3
- Google Cloud no longer sells MiniMax-M2
- Google Cloud no longer sells Llama 4 Scout
- Google Cloud no longer sells Llama 4 Maverick
- Google Cloud no longer sells Llama 3.3 70B
- Google Cloud no longer sells Kimi-K2-Thinking
- Google Cloud no longer sells Grok 4.1 Fast Reasoning
- Google Cloud no longer sells Grok 4.1 Fast Non-Reasoning
- Google Cloud no longer sells GLM-5.2
- Google Cloud no longer sells GLM-5
- Google Cloud no longer sells GLM-4.7
- Google Cloud no longer sells DeepSeek-V4.1-Flash
- Google Cloud no longer sells DeepSeek-V3.2
- Google Cloud no longer sells DeepSeek-V3.1
- Google Cloud no longer sells DeepSeek-R1 (0528)
- Google Cloud no longer sells Codestral 2
- Databricks no longer sells GPT-5.6 Luna
- Huawei Cloud no longer sells DeepSeek-V4.1-Flash (Text and image are supported)
- Together AI no longer sells thinkingmachines/Inkling
- Together AI no longer sells MiniMaxAI/MiniMax-M3
- Nebius no longer sells zai-org/GLM-5.2
- Crusoe no longer sells Qwen3 235B A22B Instruct 2507
- Crusoe no longer sells Nemotron 3.5 Super VL
- Crusoe no longer sells Nemotron 3.5 Lightning
- Crusoe no longer sells Nemotron 3 Super 120B-A12B-FP8
- Crusoe no longer sells Nemotron 3 Nano Omni 30B A3B Reasoning (Text, Image, Video)
- Crusoe no longer sells Nemotron 3 Nano 30B-A3B-FP8
- Crusoe no longer sells Gemma 4 31B-it
- Crusoe no longer sells GLM 5.3 Flash
- Crusoe no longer sells GLM 5.3
- Crusoe no longer sells DeepSeek V4.1 Flash
- Crusoe no longer sells DeepSeek V4 Flash
- Huawei Cloud no longer sells DeepSeek-V4.1-Flash
- Databricks no longer sells DeepSeek V4.1 Flash
- Azure no longer sells Fal MM H3 Max Txt to Vd 768p
- Azure no longer sells Fal MM H3 Max Txt to Vd 480p
- Azure no longer sells Fal MM H3 Max Txt to Vd 1080p
- Azure no longer sells Fal MM H3 Max Turbo Txt to Vd 768p
- Azure no longer sells Fal MM H3 Max Turbo Txt to Vd 480p
- Azure no longer sells Fal MM H3 Max Turbo Txt to Vd 1080p
- Azure no longer sells Fal MM H3 Max Turbo Img to Vd 768p
- Azure no longer sells Fal MM H3 Max Turbo Img to Vd 480p
- Azure no longer sells Fal MM H3 Max Turbo Img to Vd 1080p
- Azure no longer sells Fal MM H3 Max Ref to Vd 768p
- Azure no longer sells Fal MM H3 Max Ref to Vd 480p
- Azure no longer sells Fal MM H3 Max Ref to Vd 1080p
- Azure no longer sells Fal MM H3 Max Img to Vd 768p
- Azure no longer sells Fal MM H3 Max Img to Vd 480p
- Azure no longer sells Fal MM H3 Max Img to Vd 1080p
- Azure no longer sells Fal LTX 2.5 Pro Txt to Vd 720p
- Azure no longer sells Fal LTX 2.5 Pro Img to Vd 720p
- Azure no longer sells Fal LTX 2.5 Pro Img to Vd 1080p
- Azure no longer sells Fal LTX 2.5 Fast Txt to Vd 720p
- Azure no longer sells Fal LTX 2.5 Fast Txt to Vd 4K
- Azure no longer sells Fal LTX 2.5 Fast Txt to Vd 1440p
- Azure no longer sells Fal LTX 2.5 Fast Txt to Vd 1080p
- Azure no longer sells Fal LTX 2.5 Fast Img to Vd 4K
- Azure no longer sells Fal LTX 2.5 Fast Img to Vd 1440p
- Azure no longer sells Fal LTX 2.5 Fast Img to Vd 1080p
- DeepInfra no longer sells thinkingmachines/Inkling
- OpenAI no longer sells o4-mini
- OpenAI no longer sells o3-pro
- OpenAI no longer sells o3-mini
- OpenAI no longer sells o3
- OpenAI no longer sells o1-pro
- OpenAI no longer sells o1
- OpenAI no longer sells gpt-5.4-nano
- OpenAI no longer sells gpt-5.4-mini
- OpenAI no longer sells gpt-5.2-pro
- OpenAI no longer sells gpt-5.2
- OpenAI no longer sells gpt-5.1
- OpenAI no longer sells gpt-5-pro
- OpenAI no longer sells gpt-5-nano
- OpenAI no longer sells gpt-5-mini
- OpenAI no longer sells gpt-5
- OpenAI no longer sells gpt-4o-mini
- OpenAI no longer sells gpt-4o-2024-05-13
- OpenAI no longer sells gpt-4o
- OpenAI no longer sells gpt-4.1-nano
- OpenAI no longer sells gpt-4.1-mini
- OpenAI no longer sells gpt-4.1
- OpenAI no longer sells gpt-4-turbo-2024-04-09
- OpenAI no longer sells gpt-4-0613
- OpenAI no longer sells gpt-3.5-turbo-instruct
- OpenAI no longer sells gpt-3.5-turbo-1106
- OpenAI no longer sells gpt-3.5-turbo-0125
- OpenAI no longer sells gpt-3.5-turbo
- OpenAI no longer sells davinci-002
- OpenAI no longer sells babbage-002
- Nebius no longer sells moonshotai/Kimi-K3
- DeepInfra no longer sells openai/gpt-oss-120b-Turbo
- Nebius no longer sells moonshotai/Kimi-K3
- DeepInfra no longer sells ibm-granite/granite-4.2-3b
- DeepInfra no longer sells ibm-granite/granite-4.2-30b
- Crusoe no longer sells Yutori n2
- Alibaba Cloud no longer sells qwen3-vl-plus-2025-09-23
- Alibaba Cloud no longer sells qwen3-vl-flash-2026-01-22
- Alibaba Cloud no longer sells qwen3-vl-8b-thinking
- Alibaba Cloud no longer sells qwen3-vl-8b-instruct
- Alibaba Cloud no longer sells qwen3-vl-32b-thinking
- Alibaba Cloud no longer sells qwen3-vl-32b-instruct
- Alibaba Cloud no longer sells qwen3-vl-30b-a3b-thinking
- Alibaba Cloud no longer sells qwen3-vl-30b-a3b-instruct
- Alibaba Cloud no longer sells qwen3-vl-235b-a22b-thinking
- Alibaba Cloud no longer sells qwen3-vl-235b-a22b-instruct
- Alibaba Cloud no longer sells qwen3-next-80b-a3b-thinking
- Alibaba Cloud no longer sells qwen3-next-80b-a3b-instruct
- Alibaba Cloud no longer sells qwen3-max-2026-01-23
- Alibaba Cloud no longer sells qwen3-max
- Alibaba Cloud no longer sells qwen3-coder-480b-a35b-instruct
- Alibaba Cloud no longer sells qwen3-coder-30b-a3b-instruct
- Alibaba Cloud no longer sells qwen3-8b
- Alibaba Cloud no longer sells qwen3-30b-a3b-thinking-2507
- Alibaba Cloud no longer sells qwen3-30b-a3b-instruct-2507
- Alibaba Cloud no longer sells qwen3-30b-a3b
- Alibaba Cloud no longer sells qwen3-235b-a22b-thinking-2507
- Alibaba Cloud no longer sells qwen3-235b-a22b
- Alibaba Cloud no longer sells qwen3-14b
- Alibaba Cloud no longer sells qwen-plus-2025-09-11
- Alibaba Cloud no longer sells qwen-plus-2025-07-28
- Alibaba Cloud no longer sells deepseek-v4-flash
- Alibaba Cloud no longer sells qwq-plus
- Alibaba Cloud no longer sells qwen3.6-max-preview
- Alibaba Cloud no longer sells qwen3-vl-plus-2025-12-19
- Alibaba Cloud no longer sells qwen3-vl-flash-2025-10-15
- Alibaba Cloud no longer sells qwen3-vl-flash
- Alibaba Cloud no longer sells qwen3-omni-flash-realtime-2025-12-01
- Alibaba Cloud no longer sells qwen3-omni-flash-realtime
- Alibaba Cloud no longer sells qwen3-omni-flash-2025-12-01
- Alibaba Cloud no longer sells qwen3-omni-flash-2025-09-15
- Alibaba Cloud no longer sells qwen3-omni-30b-a3b-captioner
- Alibaba Cloud no longer sells qwen3-max-preview
- Alibaba Cloud no longer sells qwen3-max-2025-09-23
- Alibaba Cloud no longer sells qwen3-coder-plus-2025-09-23
- Alibaba Cloud no longer sells qwen3-coder-plus-2025-07-22
- Alibaba Cloud no longer sells qwen3-coder-plus
- Alibaba Cloud no longer sells qwen3-coder-next
- Alibaba Cloud no longer sells qwen3-32b
- Alibaba Cloud no longer sells qwen3-235b-a22b-instruct-2507
- Alibaba Cloud no longer sells qwen2.5-omni-7b
- Alibaba Cloud no longer sells qwen-vl-plus
- Alibaba Cloud no longer sells qwen-vl-max
- Alibaba Cloud no longer sells qwen-turbo
- Alibaba Cloud no longer sells qwen-omni-turbo-realtime-latest
- Alibaba Cloud no longer sells qwen-omni-turbo-realtime-2025-05-08
- Alibaba Cloud no longer sells qwen-omni-turbo-realtime
- Alibaba Cloud no longer sells qwen-omni-turbo-latest
- Alibaba Cloud no longer sells qwen-omni-turbo-2025-03-26
- Alibaba Cloud no longer sells qwen-omni-turbo
- Alibaba Cloud no longer sells qwen-mt-turbo
- Alibaba Cloud no longer sells qvq-max
- Alibaba Cloud no longer sells deepseek-v3.2
- AWS no longer sells TwelveLabs Pegasus 1.2 (Amazon Bedrock Edition)
- AWS no longer sells TwelveLabs Marengo Embed 3.0 (Amazon Bedrock Edition)
- AWS no longer sells TwelveLabs Marengo Embed 2.7 (Amazon Bedrock Edition)
- AWS no longer sells Stable Image Ultra (Amazon Bedrock Edition)
- AWS no longer sells Stable Image Core (Amazon Bedrock Edition)
- AWS no longer sells Stable Diffusion 3.5 Large v1.0 (Amazon Bedrock Edition)
- AWS no longer sells Stable Diffusion 3 Large v1.0 (Amazon Bedrock Edition)
- AWS no longer sells Palmyra X5 (Amazon Bedrock Edition)
- AWS no longer sells Palmyra X4 (Amazon Bedrock Edition)
- AWS no longer sells OpenAI GPT-6 Astra (Amazon Bedrock Edition)
- AWS no longer sells Meta Llama 2 Chat 70B (Amazon Bedrock Edition)
- AWS no longer sells Meta Llama 2 Chat 13B (Amazon Bedrock Edition)
- AWS no longer sells Luma Ray2 (Amazon Bedrock Edition)
- AWS no longer sells Jurassic-2 Ultra (Amazon Bedrock Edition)
- AWS no longer sells Jurassic-2 Mid (Amazon Bedrock Edition)
- AWS no longer sells Jamba-Instruct (Amazon Bedrock Edition)
- AWS no longer sells Jamba 1.5 Mini (Amazon Bedrock Edition)
- AWS no longer sells Jamba 1.5 Large (Amazon Bedrock Edition)
- AWS no longer sells Cohere Rerank v3.5 (Amazon Bedrock Edition)
- AWS no longer sells Cohere Generate Model - Command-Light (Amazon Bedrock Edition)
- AWS no longer sells Cohere Generate Model - Command (Amazon Bedrock Edition)
- AWS no longer sells Cohere Embed Model 3 - Multilingual (Amazon Bedrock Edition)
- AWS no longer sells Cohere Embed 4 Model (Amazon Bedrock Edition)
- AWS no longer sells Cohere Embed 3 Model - English (Amazon Bedrock Edition)
- AWS no longer sells Cohere Command R+ (Amazon Bedrock Edition)
- AWS no longer sells Cohere Command R (Amazon Bedrock Edition)
- AWS no longer sells Claude Sonnet 5.5 (Amazon Bedrock Edition)
- AWS no longer sells Claude Sonnet 5 (Amazon Bedrock Edition)
- AWS no longer sells Claude Sonnet 4.6 (Amazon Bedrock Edition)
- AWS no longer sells Claude Sonnet 4.5 (Amazon Bedrock Edition)
- AWS no longer sells Claude Sonnet 4 (Amazon Bedrock Edition)
- AWS no longer sells Claude Opus 5.5 (Amazon Bedrock Edition)
- AWS no longer sells Claude Opus 5 (Amazon Bedrock Edition)
- AWS no longer sells Claude Opus 4.8 (Amazon Bedrock Edition)
- AWS no longer sells Claude Opus 4.7 (Amazon Bedrock Edition)
- AWS no longer sells Claude Opus 4.6 (Amazon Bedrock Edition)
- AWS no longer sells Claude Opus 4.5 (Amazon Bedrock Edition)
- AWS no longer sells Claude Opus 4.1 (Amazon Bedrock Edition)
- AWS no longer sells Claude Opus 4 (Amazon Bedrock Edition)
- AWS no longer sells Claude Mythos Preview (Amazon Bedrock Edition)
- AWS no longer sells Claude Mythos 5.1 (Amazon Bedrock Edition)
- AWS no longer sells Claude Mythos 5 (Amazon Bedrock Edition)
- AWS no longer sells Claude Instant (Amazon Bedrock Edition)
- AWS no longer sells Claude Instant (100K) (Amazon Bedrock Edition)
- AWS no longer sells Claude Haiku 5.5 (Amazon Bedrock Edition)
- AWS no longer sells Claude Haiku 4.5 (Amazon Bedrock Edition)
- AWS no longer sells Claude Fable 5.1 (Amazon Bedrock Edition)
- AWS no longer sells Claude Fable 5 (Amazon Bedrock Edition)
- AWS no longer sells Claude 3.7 Sonnet (Amazon Bedrock Edition)
- AWS no longer sells Claude 3.5 Sonnet v2 (Amazon Bedrock Edition)
- AWS no longer sells Claude 3.5 Sonnet (Amazon Bedrock Edition)
- AWS no longer sells Claude 3.5 Haiku (Amazon Bedrock Edition)
- AWS no longer sells Claude 3 Sonnet (Amazon Bedrock Edition)
- AWS no longer sells Claude 3 Opus (Amazon Bedrock Edition)
- AWS no longer sells Claude 3 Haiku (Amazon Bedrock Edition)
- AWS no longer sells Claude (Amazon Bedrock Edition)
- AWS no longer sells Claude (100K) (Amazon Bedrock Edition)
- Core42 no longer sells k2-horizon-375b
- Azure retires gpt-5.1 (2025-11-13) on 2027-07-13
- Azure retires gpt-5-mini (2025-08-07) on 2027-07-13
- Azure retires gpt-4o (2024-05-13) on 2027-01-14; replacement: gpt-5.1 (2025-11-13)
- Mistral retires voxtral-mini-transcribe-realtime-2602 on 2027-04-30; replacement: Voxtral Mini Transcribe Realtime 3
- Mistral retires voxtral-mini-2602 on 2027-04-30; replacement: Voxtral Mini Transcribe 3
- Mistral retires mistral-large-2512 on 2027-04-30; replacement: Mistral Large 4
- Mistral retires voxtral-mini-tts-2603 on 2027-04-30; replacement: Voxtral TTS 3
- Mistral retires mistral-medium-3-5 on 2027-04-30; replacement: Mistral Large 4
- DeepInfra retires KoboldAI/LLaMA2-13B-Tiefighter on 2026-10-29; replacement: Sao10K/L3-8B-Lunaris-v1-Turbo
- DeepInfra retires Austism/chronos-hermes-13b-v2 on 2026-10-29; replacement: Sao10K/L3-8B-Lunaris-v1-Turbo
- DeepInfra retires Gryphe/MythoMax-L2-13b-turbo on 2026-10-29; replacement: Sao10K/L3-8B-Lunaris-v1-Turbo
- DeepInfra retires stepfun-ai/Step-5-Preview on 2026-10-29; replacement: zai-org/GLM-5.3-Flash
- DeepInfra retires stepfun-ai/Step-5-Preview on 2026-11-04; replacement: zai-org/GLM-5.3-Flash
- DeepInfra retires meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo on 2026-11-03; replacement: meta-llama/Llama-3.3-70B-Instruct-Turbo
- Azure retires claude-mythos-5-1 (1) on 2027-09-01
- Azure retires Codestral-2501 (2) on 2027-02-15
- Azure retires Llama-4-Scout-17B-16E-Instruct on 2027-01-31
- Azure retires grok-code-fast-1 (1) on 2027-01-31
- Azure retires grok-4 (1) on 2027-01-31
- Azure retires FLUX.2-pro (1) on 2027-01-20
- Azure retires FLUX.2-flex (1) on 2027-01-20
- Azure retires FLUX.1-Kontext-pro (1) on 2027-01-05
- Azure retires FLUX-1.1-pro (1) on 2027-01-05
- Azure retires gpt-realtime-2 (2026-05-06) on 2027-01-20; replacement: gpt-realtime-2.1
- OpenAI retires o3-pro on 2026-12-11; replacement: gpt-5.6-sol (reasoning.mode: pro)
- OpenAI retires o3 on 2026-12-11; replacement: gpt-5.6-sol
- OpenAI retires gpt-5-pro on 2026-12-11; replacement: gpt-5.6-sol (reasoning.mode: pro)
- OpenAI retires gpt-5-nano on 2026-12-11; replacement: gpt-5.6-luna
- OpenAI retires gpt-5-mini on 2026-12-11; replacement: gpt-5.6-terra
- OpenAI retires gpt-5 on 2026-12-11; replacement: gpt-5.6-sol
- OpenAI retires gpt-4o-mini-tts on 2027-01-06; replacement: gpt-realtime-2.1-mini
- Azure retires gpt-oss-120b (1) on 2027-09-15
- Azure retires Kimi-K2.7-Code (2026-06-12) on 2027-04-16
- Azure retires DeepSeek-V3.2-Speciale (1) on 2027-11-01; replacement: DeepSeek-V4-Pro
- Azure retires DeepSeek-V3.2 (1) on 2027-11-01; replacement: DeepSeek-V4.1-Flash
- Azure retires embed-v-4-0 (1) on 2027-06-01; replacement: Coher-embed-v5-fast, Cohere-embed-v5-pro
- Azure retires cohere-command-a (1) on 2027-02-15; replacement: coherelabs-command-a-plus-05-2026-fp81, coherelabs-command-a-plus-05-2026-bf161, coherelabs-command-a-plus-05-2026-w4a41
- Azure retires Cohere-rerank-v4.0-pro (1) on 2027-09-13; replacement: Cohere-rerank-v5-pro
- Azure retires Cohere-rerank-v4.0-fast (1) on 2027-09-13; replacement: Cohere-rerank-v5-fast
- DeepInfra retires Gryphe/MythoMax-L2-13b on 2026-10-29; replacement: Sao10K/L3-8B-Lunaris-v1-Turbo
- Azure retires mistral-small-2503 (1) on 2027-02-15; replacement: mistralai-mistral-small-4-119b-2603
- Azure retires Cohere-embed-v3-multilingual (1) on 2027-02-15; replacement: Cohere-Embed-V5-Fast, Cohere-Embed-V5-Pro
- Azure retires Cohere-embed-v3-english (1) on 2027-02-15; replacement: Cohere-Embed-V5-Fast, Cohere-Embed-V5-Pro
- Azure retires Llama-4-Maverick-17B-128E-Instruct-FP8 on 2026-11-19
- Azure retires Llama-3.3-70B-Instruct on 2027-01-31
- Azure retires grok-4-1-fast-reasoning (1) on 2027-05-19
- Azure retires grok-4-1-fast-non-reasoning (1) on 2027-05-19
- Azure retires Phi-4-reasoning on 2027-01-17; replacement: microsoft--phi-4 (Foundry Managed Compute)
- Azure retires Phi-4-multimodal-instruct on 2027-01-17; replacement: microsoft--phi-4 (Foundry Managed Compute)
- Azure retires Phi-4-mini-reasoning on 2027-01-17; replacement: microsoft--phi-4 (Foundry Managed Compute)
- Azure retires Phi-4-mini-instruct on 2027-01-17; replacement: microsoft--phi-4 (Foundry Managed Compute)
- Azure retires Phi-4 on 2027-01-17; replacement: microsoft--phi-4 (Foundry Managed Compute)
- AWS retires Pegasus v1.2 on 2027-03-22
- AWS retires Marengo Embed 3.0 on 2027-03-22
- DeepInfra retires NousResearch/Hermes-3-Llama-3.1-70B on 2026-11-02; replacement: deepseek-ai/DeepSeek-V4.1-Flash-1017
- DeepInfra retires NousResearch/Hermes-3-Llama-3.1-405B on 2026-11-02; replacement: deepseek-ai/DeepSeek-V4.1-Flash-1017
- DeepInfra retires openai/gpt-oss-120b-Turbo on 2026-10-20; replacement: openai/gpt-oss-120b
- AWS retires Palmyra X5 on 2027-03-21
- AWS retires Palmyra X4 on 2027-03-09
- OCI retires Cohere Embed Multilingual Light 3 on 2026-10-21; replacement: Cohere Embed 4
- OCI retires Cohere Embed English Light 3 on 2026-10-21; replacement: Cohere Embed 4
- OCI retires Cohere Embed Multilingual 3 on 2026-10-21; replacement: Cohere Embed 4
- OCI retires Cohere Embed English 3 on 2026-10-21; replacement: Cohere Embed 4
- OCI retires Cohere Embed Multilingual Image 3 on 2026-10-21; replacement: Cohere Embed 4
- OCI retires Cohere Embed English Light Image 3 on 2026-10-21; replacement: Cohere Embed 4
- OCI retires Cohere Embed English Image 3 on 2026-10-21; replacement: Cohere Embed 4
- OCI retires Cohere Embed Multilingual Light 3 on 2026-10-21; replacement: Cohere Embed 4
- OCI retires Cohere Embed English Light 3 on 2026-10-21; replacement: Cohere Embed 4
- OCI retires Cohere Embed Multilingual 3 on 2026-10-21; replacement: Cohere Embed 4
- OCI retires Cohere Embed English 3 on 2026-10-21; replacement: Cohere Embed 4
- OCI retires Cohere Embed Multilingual Image 3 on 2026-10-21; replacement: Cohere Embed 4
- Azure retires mistral-small-2503 (1) on 2026-12-09; replacement: mistral-small-4-119b-2603
- Azure retires mistral-medium-2505 (1) on 2026-12-09; replacement: mistral-medium-3-5
- Azure retires Ministral-3B (1) on 2026-12-09; replacement: ministral-ministral-4b-instruct
- Azure retires claude-opus-5-5-fast (2) on 2027-01-15
- AWS retires Claude 3.7 Sonnet on 2026-11-13
- AWS retires Claude 3.5 Sonnet v2 on 2026-11-13
- AWS retires Claude 3.5 Sonnet on 2026-11-13
- AWS retires Claude 3 Sonnet on 2026-11-13
- AWS retires Claude 3 Haiku on 2026-11-13
- AWS retires Claude 3.7 Sonnet on 2026-10-17
- AWS retires Claude 3.5 Sonnet v2 on 2026-10-17
- AWS retires Claude 3.5 Sonnet on 2026-10-17
- AWS retires Claude 3 Sonnet on 2026-10-17
- AWS retires Claude 3 Haiku on 2026-10-17
- Azure retires gpt-chat-latest (2026-10-07) on 2027-05-18
- Azure retires gpt-chat-latest (2026-08-06) on 2027-01-12
- Azure retires gpt-chat-latest (2026-06-24) on 2026-11-17
- Azure retires gpt-chat-latest (2026-05-28) on 2026-12-08
- Azure retires gpt-chat-latest (2026-05-05) on 2026-11-17
- OpenAI retires GPT-4.1 mini (gpt-4.1-mini-2025-04-14) on 2027-01-14; replacement: gpt-6-luna
- OpenAI retires GPT-4.1 (gpt-4.1-2025-04-14) on 2027-01-14; replacement: gpt-6-luna
- OpenAI retires GPT-4o mini (gpt-4o-mini-2024-07-18) on 2027-01-14; replacement: gpt-6-luna
- OpenAI retires GPT-4o (gpt-4o-2024-08-06 on 2027-01-14; replacement: gpt-6-luna
- AWS retires GLM 4.7 Flash on 2027-03-30
- AWS retires GLM 4.7 on 2027-03-30
- AWS retires Qwen3 Coder 480B A35B Instruct on 2027-03-30
- AWS retires Qwen3-Coder-30B-A3B-Instruct on 2027-03-30
- AWS retires NVIDIA Nemotron Nano 9B v2 on 2027-03-30
- AWS retires NVIDIA Nemotron Nano 12B v2 VL BF16 on 2027-03-30
- AWS retires Kimi K2 Thinking on 2027-03-30
- AWS retires Voxtral Small 24B 2507 on 2027-03-30
- AWS retires Voxtral Mini 3B 2507 on 2027-03-30
- AWS retires Mixtral 8x7B Instruct on 2027-03-30
- AWS retires Mistral 7B Instruct on 2027-03-30
- AWS retires Devstral 2 123B on 2027-03-30
- AWS retires MiniMax M2.1 on 2027-03-30
- AWS retires MiniMax M2 on 2027-03-30
- AWS retires Llama 4 Scout 17B Instruct on 2027-03-30
- AWS retires Llama 4 Maverick 17B Instruct on 2027-03-30
- AWS retires Llama 3.3 70B Instruct on 2027-03-30
- AWS retires Llama 3.1 8B Instruct on 2027-03-30
- AWS retires Llama 3.1 70B Instruct on 2027-03-30
- AWS retires Llama 3 8B Instruct on 2027-03-30
- AWS retires Llama 3 70B Instruct on 2027-03-30
- AWS retires Gemma 3 4B IT on 2027-03-30
- AWS retires Gemma 3 27B PT on 2027-03-30
- AWS retires Gemma 3 12B IT on 2027-03-30
- AWS retires DeepSeek-V3.1 on 2027-03-30
- AWS retires DeepSeek-R1 on 2027-03-30
- AWS retires Claude Sonnet 4.5 on 2027-04-08
- Azure retires gpt-4.1-nano (2025-04-14) on 2027-04-14
- Google Cloud retires gemini-3.6-flash on 2026-11-19; replacement: gemini-3.8-flash
- Google Cloud retires gemini-3.7-flash on 2027-01-28; replacement: gemini-3.8-flash
New GPU instances, by cloud
Alibaba Cloud 37, Core42 4, Azure 4
New regions, by cloud
Azure 92, AWS 31, Nebius 2, Fireworks AI 1
The clouds' own announcements
- AWS: Amazon Bedrock expands API support and introduces Cross-Region Inferencing for OpenAI models in AWS GovCloud (US)
- AWS: Daybreak Red from OpenAI now available to eligible customers on Amazon Bedrock in AWS GovCloud (US-East)
- OCI: New tutorial: Search Documents with Vector Stores and OCI Generative AI API Keys
- AWS: Amazon SageMaker AI inference now supports GPU partitioning for real-time endpoints
- OpenAI: Added opt-in experimental time awareness.
- OpenAI: You can now update the tools in an existing session in the Agents API public beta.
- OpenAI: Added automatic review for eligible Agents API projects.
- Databricks: DeepSeek V4 Pro (0813) is retired
- Databricks: Moonshot AI Kimi K2.7 is retired
- Anthropic: We’ve deprecated fast mode for Claude Opus 4.8, with removal on November 30, 2026.
- Snowflake: Oct 30, 2026: Cortex Training (Preview)
- Snowflake: Cortex Training (Preview)
- OpenAI: Released GPT-6 Rosalind, bringing specialized biological reasoning with scientific tool use to complex life sciences research.
- Telnyx: gpt-oss-120b Now Available on Telnyx Inference
- AWS: Amazon Bedrock AgentCore CLI version 1.0 is now available
- Google Cloud: Fixed: CodeMender updates (v0.17.0)
- Google Cloud: Change: Use a single regional Agent Registry instance for Gemini Enterprise apps
- Snowflake: Oct 28, 2026: OpenAI GPT-5.6 Luna, GPT-5.6 Terra, and GPT-6 Luna in AI Functions (Preview)
- Snowflake: OpenAI GPT-5.6 Luna, GPT-5.6 Terra, and GPT-6 Luna in AI Functions (Preview)
- OpenAI: Eligible ChatGPT Pro and Business subscribers can now claim up to $500 per month in free API credits as a subscription benefit.
- Snowflake: Oct 28, 2026: Cortex AI usage in Native App usage history
- Snowflake: Cortex AI usage in Native App usage history
- AWS: Amazon ECS now supports GPU sharing with NVIDIA Multi-Process Service (MPS)
- Azure: [Launched] Generally Available: Cost management and tagging for Claude models in Microsoft Foundry
- AWS: Amazon Bedrock now offers Canada cross-Region inference for Claude Haiku 5.5
- Google Cloud: Feature: xAI's Grok 4.7 is generally available
- Google Cloud: Feature: Gemini Robotics ER 2 (Preview)
- Snowflake: Oct 27, 2026: OpenAI GPT-6.1 Terra (Preview)
- Snowflake: OpenAI GPT-6.1 Terra (Preview)
- Huawei Cloud: Model rollout: DeepSeek‑V4.1‑Flash-1017 is now available, featuring a 1M‑token context window along with deep thinking, function calling, continuation of Prefix, and prefix caching.
- Snowflake: October 27, 2026: Cortex Code skills for interactive analytics
- Snowflake: Cortex Code skills for interactive analytics
- Alibaba Cloud: qwen3.8-flash: Qwen3.8-Flash is the latest multimodal model from the Qwen family, combining powerful reasoning and generation with remarkable speed.
- Databricks: OpenAI GPT-6.1 Terra is now available on Unity Gateway
- Databricks: OpenAI GPT-6.1 Terra is now available on Unity Gateway
- AWS: OpenAI GPT-6.1 Terra is now generally available on Amazon Bedrock
- Google Cloud: Feature: Semantic governance policies support more model providers
- Google Cloud: Feature: Semantic governance policies support more model providers
- Google Cloud: Feature: Cloud Storage output support for Gemini image generation models
- Huawei Cloud: Model rollout: DeepSeek‑V4.1‑Flash-1017 is now available, featuring a 1M‑token context window along with deep thinking, function calling, continuation of Prefix, and prefix caching.
- Huawei Cloud: New guide to improving token cache hit rates: Explains high cache hit rate scenarios, how caching works, best practices for improving cache hit rates, and case studies of low cache hit rates, and helps you improve model…
- Huawei Cloud: New integration guides for TRAE and ZCode: Added instructions for integrating MaaS model APIs with TRAE and ZCode.
- Alibaba Cloud: glm-5.3: glm-5.3 is now available with the US deployment scope in the US (Virginia) region, supporting text generation, deep thinking, Function Calling, structured output, and context…
- Alibaba Cloud: `glm-5.3`: glm-5.3 is now available with the US deployment scope in the US (Virginia) region, supporting text generation, deep thinking, Function Calling, structured output, and context…
- OpenAI: See the GPT-6 guide for prompting and migration guidance.
- OpenAI: Released GPT-6.1 Terra, a fast, efficient model for visual reasoning, precise computer use, and clear, natural writing.
- AWS: Web Search on AWS Bedrock now supports domain filtering and user location controls
- Snowflake: Oct 23, 2026: Responses API for the Cortex REST API (Preview)
- Snowflake: Responses API for the Cortex REST API (Preview)
- AWS: Amazon SageMaker AI announces training session APIs for serverless model customization
- OCI: Validate GPU cluster health with Cluster Health Checks
- AWS: AgentCore Gateway now supports Bedrock InvokeModel APIs and model fallback
- AWS: Amazon SageMaker AI now supports serverless model customization for Muse Glimmer, Qwen 3.8 and Ministral 3 models
- Snowflake: Oct 23, 2026: Retrieve Cortex Search (Public Preview)
- Snowflake: Retrieve Cortex Search (Public Preview)
- Anthropic: We’ve launched Claude Models on Google Cloud (Anthropic managed), the Claude API operated by Anthropic and reached through your Google Cloud project.
- Anthropic: We've launched Claude Models on Google Cloud (Anthropic managed), the Claude API operated by Anthropic and reached through your Google Cloud project.
- OpenAI: The Decisions API is now generally available with gpt-6-luna.
- OCI: OCI Kubernetes Engine (OKE) support for GPU maintenance event repair
- OCI: Create Generative AI applications for public content
- OCI: Add file storage to Generative AI applications
- AWS: Amazon SageMaker AI now supports custom sequence lengths for serverless model customization jobs
- Anthropic: We’ve released API version 2026-10-21.
- Anthropic: We've released API version 2026-10-21.
- Cloudflare: Agents, Workers AI, AI Gateway - createAI replaces workers-ai-provider and ai-gateway-provider
- AWS: Structured outputs for Claude models available on Amazon Bedrock
- Mistral: Voxtral 3
- OCI: OCI Generative AI expands on-demand model access to 30 commercial regions
- OCI: Import NVIDIA Nemotron 3.5 Super VL NVFP4 into OCI Generative AI
- OCI: Import Alibaba Qwen3.6-35B-A3B-FP8 into OCI Generative AI
- OCI: OCI Generative AI adds CPU-only access in 30 commercial regions
- OCI: Use Sandboxes in OCI Generative AI
- AWS: Amazon Bedrock Advanced Prompt Optimization now supports prompt shortening
- AWS: Amazon Bedrock Guardrails now supports Standard tier for contextual grounding checks
- Snowflake: Oct 20, 2026: Cortex Fine-tuning: New model training no longer available in AWS US East 1
- Snowflake: Cortex Fine-tuning: New model training no longer available in AWS US East 1
- OpenAI: You can now choose 12-hour prompt caching in the API. 12-hour prompt caching keeps shared instructions, tools, and context cached longer to benefit from lower cached-input…
- AWS: Grok 4.7 is now available on Amazon Bedrock in AWS GovCloud (US)
- Google Cloud: Feature: Supervised fine-tuning for Gemini 3 models is available in General Availability
- Snowflake: Oct 19, 2026: Tools for Cortex AI Gateway (Preview)
- Snowflake: Tools for Cortex AI Gateway (Preview)
- Fireworks AI: New models on Serverless
- Fireworks AI: Recommended migrations
- Fireworks AI: Action required
- Fireworks AI: Upcoming Serverless deprecation: GPT OSS 120B, Inkling, Nemotron, and US endpoints for MiniMax M3 and GLM 5.3 Flash
- Databricks: DeepSeek V4.1 Flash (1017) is now available on Unity Gateway
- Databricks: DeepSeek V4.1 Flash is scheduled for retirement
- Databricks: DeepSeek V4.1 Flash (1017) is now available on Unity Gateway
- Snowflake: Oct 19, 2026: Cortex Tools for Cortex AI Gateway (Preview)
- Snowflake: Cortex Tools for Cortex AI Gateway (Preview)
- Cloudflare: Workers AI - DeepSeek V4.1 Flash 1017 now available on Workers AI
- Cloudflare: Workers AI - Introducing Clairic: an extraction model for typed values, now on Workers AI
- Together AI: New serverless models: deepseek-ai/DeepSeek-V4.1-Flash-1017
- Fireworks AI: Expanded Small serverless rate-limit tier
- Alibaba Cloud: deepseek-v4.1-flash-1017: DeepSeek-V4.1-Flash-1017 is the latest iteration of V4.1-Flash, trained with self-feedback-based reinforcement learning to deliver substantial gains over its predecessor across…
- Alibaba Cloud: `deepseek-v4.1-flash-1017`: DeepSeek-V4.1-Flash-1017 is the latest iteration of V4.1-Flash, trained with self-feedback-based reinforcement learning to deliver substantial gains over its predecessor across…
- AWS: Nemotron 3.5 Super VL model now available on Amazon SageMaker JumpStart
- AWS: Beam model by Reflection is now available on Amazon SageMaker JumpStart and AWS Trainium
- OpenAI: Added Ultrafast mode to the Agents API.
- Snowflake: Oct 16, 2026: Grok 4.7 and GLM 5.3 models (Preview)
- Snowflake: Grok 4.7 and GLM 5.3 models (Preview)
- Snowflake: Oct 16, 2026: Cortex AI Gateway on Azure, Google Cloud, and AWS New Zealand
- Snowflake: Cortex AI Gateway on Azure, Google Cloud, and AWS New Zealand
- Together AI: New serverless models: reflection/Beam-1.0-501B-A23B
- OpenAI: Added nine natural voices for GPT-Live: Nuri, Juni, and Flitz (German); Brise and Sillage (French); Nira (Hindi); Harema and Shida (Japanese); and Noeul (Korean).
- OpenAI: Released prompt-to-voice for all API customers.
- AWS: Web Search on Amazon Bedrock AgentCore expands to AWS GovCloud and 3 Regions
- AWS: Muse Spark Open model now available on Amazon SageMaker JumpStart
- OCI: Import NVIDIA Nemotron 3 Ultra FP8 into OCI Generative AI
- OCI: Expanded hardware support for imported models in OCI Generative AI
- OCI: Import Xiaomi MiMo V2.6 Pro RL into OCI Generative AI
- OCI: Import Alibaba Qwen3.8-Flash-Next into OCI Generative AI
- Mistral: Introducing Timestral
- OpenAI: Added native Decisions integration in GPT-Live.
- Cloudflare: Workers AI, Realtime - Pipecat PhoneLLM Mini now available on Workers AI
- Anthropic: Fast mode for Claude Opus 5.5 is now available on Claude in Microsoft Foundry for deployments hosted on Azure.
- Google Cloud: Feature: DeepSeek-V4.1-Flash is available in Public Preview
- Google Cloud: Fixed: CodeMender updates (v0.15.0)
- Google Cloud: Feature: Priority PayGo support for Gemini Nano Banana 2.1
- AWS: Amazon Bedrock AgentCore Memory now supports configurable time-to-live for long-term memory
- Google Cloud: Feature: Gemini 4 Argon in CodeMender (Preview)
- Snowflake: Oct 13, 2026: Gemini 4 Argon model in AI Functions (Preview)
- Snowflake: Gemini 4 Argon model in AI Functions (Preview)
- Databricks: Google Gemini 4 Argon is now available on Unity Gateway
- Databricks: Google Gemini 4 Argon is now available on Unity Gateway
- AWS: Announcing consumption-based pricing for short-term memory in Amazon Bedrock AgentCore
- AWS: Amazon Bedrock now supports reasoning summaries for OpenAI models
- AWS: TwelveLabs Pegasus 1.5 model now available on Amazon Bedrock
- Cloudflare: Workers AI - Clef-omni adds audio and video input, Clef-flash is now cheaper, and Clef is faster
- AWS: OpenAI GPT-6.1 Sol now supports Ultrafast mode on Amazon Bedrock
- AWS: AWS Cost Explorer, Budgets, and Dashboards now support Amazon Bedrock product attributes
- Alibaba Cloud: qwen3.7-flash: qwen3.7-flash is now available with the US deployment scope in the US (Virginia) region, supporting text generation, deep thinking, visual understanding, Function Calling,…
- Alibaba Cloud: qwen3.8-max: qwen3.8-max is now available with the US deployment scope in the US (Virginia) region, supporting text generation, deep thinking, visual understanding, Function Calling,…
- OpenAI: Added Ultrafast mode for GPT-6.1 Sol in the Responses API.
- OCI: Use identity domain browser session authentication for hosted applications in OCI Generative AI
- Google Cloud: Feature: Anthropic's Claude Haiku 5.5
- Anthropic: We’ve launched Claude Haiku 5.5 (claude-haiku-5-5), our most capable model tuned for high-volume and latency-sensitive work.
- Anthropic: We’ve lowered the price of prompt cache reads on Claude Sonnet 5.5 from $0.20 USD to $0.10 USD per million tokens: 0.05x the base input price instead of 0.1x.
- Snowflake: Oct 07, 2026: Claude Haiku 5.5 model in the Cortex REST API (Preview)
- Snowflake: Oct 7, 2026: Snowflake Decision with AI_COMPLETE (Private Preview)
- Snowflake: Claude Haiku 5.5 model in the Cortex REST API (Preview)
- Snowflake: Snowflake Decision with AI_COMPLETE (Private Preview)
- Fireworks AI: Doubled serverless rate limits for Small models
- Databricks: Anthropic Claude Haiku 5.5 is now available on Unity Gateway
- Databricks: Anthropic Claude Haiku 5.5 is now available on Unity Gateway
- Databricks: DeepSeek V4.1 Flash supports priority pay-per-token and reserved provisioned throughput
- Anthropic: We've launched Claude Haiku 5.5 (claude-haiku-5-5), our most capable model tuned for high-volume and latency-sensitive work.
- Anthropic: We've lowered the price of prompt cache reads on Claude Sonnet 5.5 from $0.20 USD to $0.10 USD per million tokens: 0.05x the base input price instead of 0.1x.
- OpenAI: Updated the chat-latest snapshot, which points to the latest model available in ChatGPT for Plus, Pro, Business and Enterprise users.
- Mistral: Introducing Mistral Large 4
- Google Cloud: Feature: Gemini Nano Banana 2.1
- Snowflake: Oct 06, 2026: Kimi K3 model (Preview)
- Snowflake: Kimi K3 model (Preview)
- Databricks: DeepSeek V4 Flash (0731) is scheduled for retirement
- OpenAI: Simplified API usage tiers from five to three: Build, Launch, and Grow.
- OpenAI: Released the Decisions API in beta with gpt-6-luna.
- AWS: GLM 5.3 by Z.ai is now generally available on Amazon Bedrock
- AWS: Announcing Amazon Nova 2.5 Sonic with improved reasoning for voice agents
- Snowflake: Oct 05, 2026: OpenAI GPT-6.1 Sol model (Preview)
- Snowflake: OpenAI GPT-6.1 Sol model (Preview)
- Alibaba Cloud: qwen3.8-flash: qwen3.8-flash is now available with the US deployment scope in the US (Virginia) region, supporting text generation, deep thinking, visual understanding, Function Calling,…
- OCI: OCI Kubernetes Engine (OKE) support for GPU host repair
- Databricks: Google Gemini 2.5 Flash is retired
- Databricks: Google Gemini 2.5 Pro is retired