Latest AI model releases.
Language and speech model announcements, with the official source one click away. Dates reflect the provider’s announcement; availability can vary by platform and region.
moonshotai/kimi-k3.1-pro-highspeed
Impossibl announced this model on 2026-10-23. ai api.
MiniMaxAI/MiniMax-M3.1-Flash
Together AI announced this model on 2026-10-23.
moonshotai/kimi-k3.1-pro
Impossibl announced this model on 2026-10-23. ai api.
zai-glm-5-3-flash
Mistral announced this model on 2026-10-22.
rerank-v5.0-pro
Cohere announced this model on 2026-10-21.
rerank-v5.0-fast
Cohere announced this model on 2026-10-21.
voxtral-mini-transcribe-3-0
Mistral announced this model on 2026-10-20. Public Preview.
voxtral-small-transcribe-3-0
Mistral announced this model on 2026-10-20. Public Preview.
voxtral-mini-transcribe-realtime-3-0
Mistral announced this model on 2026-10-20. Public Preview.
voxtral-mini-tts-3-0
Mistral announced this model on 2026-10-20. Public Preview.
Mercury 3
Inception announced this model on 2026-10-19.
DeepSeek v4.1 Flash 1017
cf/deepseek-ai/deepseek-v4.1-flash-1017Cloudflare Workers AI announced this model on 2026-10-19.
minimaxai/minimax-m3.1-flash
Impossibl announced this model on 2026-10-19. ai api.
DeepSeek V4.1 Flash 1017
deepseek-ai/DeepSeek-V4.1-Flash-1017Together AI announced API availability for this model on 2026-10-18. serverless
DeepSeek v4.1 Flash 1017
deepseek/deepseek-v4.1-flash-1017Impossibl announced this model on 2026-10-18. ai api.
GLM-5.2
IteraCompute announced API availability for this model on 2026-10-16.
typesafe-ai/jev-1.14.0
Impossibl announced this model on 2026-10-15. gateway.
accounts/fireworks/models/muse-spark-open-nvfp4
Fireworks AI announced API availability for this model on 2026-10-15. Serverless Training
muse-spark-open
AssemblyAI announced this model on 2026-10-15. LLM Gateway.
meta-models/Muse-Spark-Open-816B-A42B
Together AI announced API availability for this model on 2026-10-15. serverless
meta/muse-spark-open
Impossibl announced this model on 2026-10-15. ai api.
cf/daily-co/phonellm-mini-alpha-2
Cloudflare Workers AI announced this model on 2026-10-15.
nvidia/NVIDIA-Nemotron-3.5-Super-VL
Together AI announced API availability for this model on 2026-10-15. serverless
meta/muse-wonder
Impossibl announced this model on 2026-10-14.
Gemini 4 argon
gemini-4-argonSnowflake Cortex announced this model on 2026-10-13. Public Preview in AI Functions with AI_COMPLETE.
Bulbul V4 Flash
Sarvam AI announced this model on 2026-10-13. India.
stepfun/step-5-preview
Impossibl announced this model on 2026-10-09. gateway.
stepfun/step-3.7-flash
Impossibl announced this model on 2026-10-09.
GPT 6 luna decisions
openai/gpt-6-luna-decisionsImpossibl announced this model on 2026-10-08. /v1/decisions.
snowflake-decision
Snowflake Cortex announced this model on 2026-10-07.
Claude Haiku 5.5
claude-haiku-5-5Snowflake Cortex announced API availability for this model on 2026-10-07. Public Preview in the Cortex REST API
mistral-large-4
Mistral announced this model on 2026-10-06.
mimo-v2.6-pro
Neuralwatt announced this model on 2026-10-06.
mimo-v2.6-pro-flex
Neuralwatt announced this model on 2026-10-06.
Gemini Nano banana 2.1
google/gemini-nano-banana-2.1Impossibl announced this model on 2026-10-06. /v1/images/generations.
mistral/mistral-large-4
Impossibl announced this model on 2026-10-06.
tencent/hy4-preview
Impossibl announced this model on 2026-10-05. joins the ai api.
xiaomi/mimo-v2.6-pro
Impossibl announced this model on 2026-10-05. ai api.
xiaomi/mimo-v2.6-flash
Impossibl announced this model on 2026-10-05.
openai-gpt-6.1-sol
Snowflake Cortex announced this model on 2026-10-05.
zai-org/GLM-5.3
Together AI announced this model on 2026-10-02. provisioned throughput.
NVIDIA Nemotron 3.5 Super VL
DigitalOcean announced API availability for this model in Oct 2026.
Beam
DigitalOcean announced this model on 2026-10-01. DigitalOcean Inference for serverless inference.
Claude Haiku 5.5
DigitalOcean announced this model on 2026-10-01.
MiniMax M3.1-Flash
DigitalOcean announced this model on 2026-10-01.
DeepSeek V4.1 Flash 1017
DigitalOcean announced API availability for this model in Oct 2026. DigitalOcean Inference
Qwen3 TTS 12Hz 1.7B CustomVoice
Qwen/Qwen3-TTS-12Hz-1.7B-CustomVoiceTogether AI announced this model on 2026-10-01. dedicated endpoints.
ocr-260930
Upstage announced this model on 2026-09-30.
document-parse-260930
Upstage announced this model on 2026-09-30.
About these model releases
Does a release mean a model is available everywhere?
No. An announcement applies to the provider and hosting route shown. API access, regions and availability on other platforms may differ. Open the official source for the provider’s terms.
Where do release dates come from?
Dates come from the provider’s announcement. A month-only date means the source does not establish an exact day; it is not a guessed release date.
How we verify notices → · Check a model’s retirement status →