Back to Blog
blog••
7

AI model ecosystem briefing — 2026-11-16

Unattended daily digest of the LLM API landscape: 4 releases, 2 new models, 28 pricing updates and 1 community signal.

1. Today at a glance

  • Models tracked: 469
  • New models: 2
  • Pricing updates: 28
  • Community signals: 1

This briefing is generated by the unattended AINode content engine from public model catalogs, GitHub releases and developer community discussions.

2. New models

Model IDProviderContextInput /1MOutput /1M
tencent/hy4-flash tencent 1.0M $0.044 $0.176
qwen/qwen4-max qwen 1M $1.50 $4.50

3. Pricing updates

  • moonshotai/kimi-k3.1-pro:batch input $0.404 → $0.381 (−6%)
  • moonshotai/kimi-k3.1-pro:batch output $6.06 → $5.71 (−6%)
  • deepseek/deepseek-v4.1-flash-1017 input $0.020 → $0.150 (+650%)
  • deepseek/deepseek-v4.1-flash-1017 output $0.108 → $0.600 (+456%)
  • deepseek/deepseek-v4.1-flash-1017:batch input $0.038 → $0.037 (−3%)
  • deepseek/deepseek-v4.1-flash-1017:batch output $0.336 → $0.326 (−3%)
  • ~deepseek/deepseek-flash-latest input $0.020 → $0.036 (+80%)
  • ~deepseek/deepseek-flash-latest output $0.108 → $0.500 (+363%)
  • deepseek/deepseek-v4.1-flash input $0.051 → $0.140 (+176%)
  • deepseek/deepseek-v4.1-flash output $0.203 → $0.420 (+107%)
  • z-ai/glm-5.3 input $0.084 → $0.070 (−17%)
  • z-ai/glm-5.3 output $3.60 → $1.40 (−61%)

4. Official & open-source releases

  • vllm-project/vllm v0.34.0:v0.34.0 View release ↗
  • ollama/ollama v0.40.17:v0.40.17
    • Improve ChatGPT CLI restore and graceful restart by @hoyyeva in #19583
    • cmd/launch: use thinking metadata for OpenCode settings by @hoyyeva in #19409
    • llm: use mmap for Vulkan integrated GPUs by @dhiltgen in #19640
    • rpreslar4765/Ollama by @rpreslar4765 in #17163
    View release ↗
  • deepseek-ai/DeepSeek-V3 v1.0.0:v1.0.0 View release ↗
  • meta-llama/llama-models v0.2.0:v0.2.0 View release ↗

5. Community pulse

6. Gateway tip

Keep one OpenAI-compatible endpoint and one key across providers. Route failover, meter tokens and bill by model without touching provider SDKs.