OpenCode Go US$5/Month (~A$7) for First Month Then US$10/Month (~A$14)

• +5·Smilesy·3d ago

  I’ve been using OpenCode Go for about a month, and overall it’s excellent value if you’re getting into vibe coding or other AI‑assisted workflows. For light or casual use, the included quota goes a long way and feels very cost‑effective. Heavy users will burn through it pretty quickly, but that’s expected with any usage‑based plan. In terms of practical output, I found the experience comparable to what I typically get from a Claude Pro subscription when using it mainly for coding tasks — especially with the Sonnet and Haiku models. For dabblers or people experimenting with AI‑assisted development, it’s a great option.

• +1·3dprintdeals·2d ago

  Great for Hermes + Kimi K2.6, Qwen3.6 Plus & Deepseek V4 Flash Model Provider Requests / 5 Hours Requests / Week Requests / Month Notes GLM-5.1 DeepInfra, Fireworks AI, Z.ai 880 2,150 4,300 Higher-cost model GLM-5 DeepInfra, Fireworks AI, Z.ai 1,150 2,880 5,750 Kimi K2.5 Moonshot AI 1,850 4,630 9,250 Kimi K2.6 Moonshot AI 1,150 2,880 5,750 MiMo-V2.5-Pro Xiaomi MiMo 1,290 3,225 6,450 MiMo-V2.5 (≤ 256K) Xiaomi MiMo 2,150 5,450 10,900 Context window ≤ 256K Qwen3.5 Plus Alibaba Cloud Model Studio 10,200 25,200 50,500 Lower-cost model Qwen3.6 Plus Alibaba Cloud Model Studio 3,300 8,200 16,300 MiniMax M2.7 MiniMax 3,400 8,500 17,000 MiniMax M2.5 MiniMax 6,300 15,900 31,800 Lower-cost model DeepSeek V4 Pro DeepSeek 3,450 8,550 17,150 DeepSeek V4 Flash DeepSeek 31,650 79,050 158,150 Lowest-cost model Usage Limits Window Included Usage 5 Hour Limit $12 of usage Weekly Limit $30 of usage Monthly Limit $60 of usage Limits are defined in dollar value. Actual request counts vary depending on model pricing and token usage patterns

• +1·Meron35·2d ago

  Ollama Cloud is by far the best value right now. $20 USD/month for a weekly limit that matches nearly perfectly with a 5 day work week. Pricing is admittedly opaque, but they charge based on compute, not tokens, so with some clever config you can delegate to models with cheaper compute. Main issue is reportedly slow/unreliable service during US peak hours, which isn't an issue in Oz. By comparison, the entire $60 of value in OpenCode Go only lasts me around half a week at best.

• +1·ClearanceX·21h ago

  Been using Go for about a month now. Started with it on OpenClaw, then moved to Hermes Agent — currently running DeepSeek V4 Pro and it's been solid. The killer feature for me is that Go gives you a proper API endpoint you can point anywhere. OpenClaw, Hermes, whatever custom setup you're running — it just works. For context: I used to jump between providers trying to find consistent quality without paying $30+/month. Codex was my go-to for a while. Go at $10/month solved that. Privacy angle is underrated too — zero-retention policy, models hosted US/EU/Singapore. Only thing worth flagging: the referral credit is for overflow usage, not a straight discount. Still a nice-to-have.

• +1·ClearanceX·18h ago

  Haven't tried DeepSeek V4 Flash yet — I set up a separate OpenCode account just for Hermes and DeepSeek V4 Pro is the sweet spot for me. Tried GLM-5.1 briefly but didn't see any real advantage over V4 Pro, if anything V4 Pro felt more consistent. One thing worth noting: OpenCode Go models are text-only, so if you're using Hermes with image inputs it won't work.

Captured from OzBargain at crawl time. For the full thread, follow the deal link.