Changelog
What changed in free models, limits, and promos.
2026-10-10
- Removed ended deals: Ox Alpha, GLM-5.3 Flash discount, Hy4 preview and Hy3 WorkBuddy windows
- Groq free plan now GPT-OSS 20B/120B and Qwen3.8 27B at 1,000 requests/day; Groq retired Llama 3.1 8B and 3.3 70B on 2026-08-16
- Llama 3.3 and 3.1 free API now via Cloudflare Workers AI; Cerebras moved to a one-time trial credit (no recurring free tier)
- Deals: Step 5 Preview (OpenCode, Kilo, Cline, to ~10-15), Ling 3.1 Flash in Cline (to 10-13), Laya on Vercel AI Gateway and Poolside Laguna :free (to 10-31)
- Gemini free tier now Gemini 3.8 Flash / 3.5 Flash-Lite (2.5 only for existing projects); Google and Mistral no longer publish fixed free limits — shown as notes instead of numbers
- Cloudflare Workers AI limit shown as 10,000 Neurons/day, not requests; NVIDIA NIM 40 RPM marked as per-model and unpublished
- Fixed Groq compare pages (1,000 requests/day, no free Llama)
- Added Qwen3.8 27B and Gemini 3.8 Flash model cards
- New guides: Groq retired Llama replacements; free models in Cline, Kilo Code and OpenCode
- K2 Horizon 7B has no working Ollama tag yet — replaced with llama.cpp fork and vLLM commands
2026-10-06
- Added Apodex 1.1 Mini, Dots3-Note Preview, Ling 3.0 Flash Sante, Ling 3.1 Flash, and Nemotron 3.5 Content Safety
- OpenRouter :free — Nex-N2.5 Pro/Mini and Ling 3.0 Flash Fin left the free roster; Nex cards marked paid
- Archived Hy3 WorkBuddy free window (ended 2026-09-30)
- Added 15 local Ollama model cards and run pages — Qwen2.5 Coder, Qwen3 Coder, DeepSeek Coder V2, Gemma 3, Phi-3, Granite 4, SmolLM2, Llama 3.2 Vision, Mistral Nemo, Command R, Hermes 3, QwQ, GPT-OSS, Devstral, OLMo 2
2026-10-05
- OpenRouter :free routes (17 models) synced on 2026-10-05
- Provider limits synced (gemini-free, groq, cloudflare-workers-ai, mistral-free, cerebras-free, nvidia-nim, openrouter-free)
2026-09-28
- OpenRouter :free routes (17 models) synced on 2026-09-28
- Provider limits synced (gemini-free, groq, cloudflare-workers-ai, mistral-free, cerebras-free, nvidia-nim, openrouter-free)
2026-09-21
- OpenRouter :free routes (22 models) synced on 2026-09-21
- Provider limits synced (openrouter-free, cerebras-free, gemini-free, cloudflare-workers-ai, groq, nvidia-nim, mistral-free)
2026-09-14
- OpenRouter :free routes (20 models) synced on 2026-09-14
- Provider limits synced (openrouter-free, cerebras-free, gemini-free, cloudflare-workers-ai, groq, nvidia-nim, mistral-free)
2026-09-10
- OpenRouter :free synced — MiniMax M3/M2.7 and GLM-5.2 free promos ended; marked paid on model cards
- Added Nex-N2.5 Pro and Nex-N2.5 Mini free API cards
- Compares — Nex-N2.5 Pro vs Qwen3, Nex-N2.5 Pro vs DeepSeek R1
- Deals — archived GLM-5.3 Flash launch discount and Hy4 WorkBuddy free window; Hy3 free still through 2026-09-30
- OpenRouter :free routes (18 :free + openrouter/free) synced on 2026-09-10
2026-09-07
- OpenRouter :free routes (19 models) synced on 2026-09-07
- Provider limits synced (openrouter-free, cerebras-free, gemini-free, cloudflare-workers-ai, groq, nvidia-nim, mistral-free)
2026-09-06
- P1 — FAQ + SEO intros on Llama 3.1/3.3 and Gemini free-api pages
- P1 — Free API stack guide (Groq + OpenRouter + Gemini)
- P2 — Model cards: Gemma 4 31B, LFM 2.5, Laguna XS 2.1, Inkling Small
- P2 — vLLM Linux install guide linked from VPS guide and Ollama vs vLLM
- SEO refresh for Ollama Windows/macOS/Docker and Cursor macOS install titles
- Added Cursor Linux, LM Studio macOS and Linux install guides
- Added NVIDIA NIM and Mistral free provider pages to API comparison
- Synced OpenRouter :free roster (Gemma 4, Nemotron 3 Ultra/Super/Lightning, MiniMax M3/M2.7, Inkling, Laguna, North Mini Code, GLM-5.2)
- Added model cards for MiniMax M3, Nemotron Ultra/Super/Lightning, Inkling, Laguna S 2.1, North Mini Code, GLM-5.3 Flash, K2 Horizon 7B, Hy4 preview
- Codestral free API path via Mistral Experiment tier
- Deals — Hy4 WorkBuddy free through 2026-09-10, Hy3 through 2026-09-30, GLM-5.3 Flash launch discount through 2026-09-09; archived Ox Alpha
- Compares — NVIDIA NIM vs Groq, Mistral vs Groq, Ollama vs vLLM
- Strengthened Llama/Gemini/Groq free-API copy and home internal links
- OpenRouter :free routes (20 models) synced on 2026-09-06
- Provider limits synced (cloudflare-workers-ai, cerebras-free)
2026-08-31
- OpenRouter :free routes (19 models) synced on 2026-08-31
- Provider limits synced (groq, gemini-free, cloudflare-workers-ai, cerebras-free, openrouter-free)
2026-08-24
- Added FAQ JSON-LD schema to install guides, Ollama pages, and VPS guide
- Added Ollama hub page with install guides and model table
- Added Ollama install guides for macOS and Docker
- Expanded Windows and Linux Ollama guides with API server steps
- Added local Ollama support for Mixtral and Llama 3.3
- Enhanced per-model Ollama pages with pull and API sections
- Expanded model catalog from 4 to 15 models
- Added Llama 3.3, Mixtral, Gemini 2.5 Flash, Gemma 4, Nemotron 3, GLM-5, Phi-4, Qwen 2.5, Llama 3.2, Codestral, LLaVA
- Added LM Studio Windows install guide
- Added Gemini vs Groq and Ollama vs llama.cpp comparisons
- Added OpenRouter vs Groq, Gemini vs OpenRouter, Ollama vs Open WebUI, Cloudflare vs Groq, and LM Studio vs llama.cpp comparisons
- Re-verified provider limits (checked_at 2026-08-24)
- OpenRouter :free routes (19 models) synced on 2026-08-24
- Updated OpenRouter :free routes — old Llama/DeepSeek/Qwen routes removed, 19 new routes active
- Added per-model free API pages with provider limits and curl examples
- Provider limits synced (groq, gemini-free, cloudflare-workers-ai, cerebras-free, openrouter-free)
2026-08-23
- Site launch with API comparison, provider limits, and model catalog
- Added Groq, OpenRouter, Gemini, Cloudflare Workers AI, Cerebras provider pages
- Ox Alpha temporary promo tracked on deals page
- OpenRouter :free routes (19 models) synced on 2026-08-23
- Provider limits synced (cerebras-free, cloudflare-workers-ai, gemini-free, groq, openrouter-free)