Filter
Exclude
Time range
-
Near
(。-`ω-)ンー Ollamaに課金してクラウドモデル使う方が いいのか それともopencodeGOに課金してクラウドモデル 使う方がいいのか どちらもHermèsagentのモデルで使うだけだけど、 ローカルなのにクラウドモデルって脳がバグるなぁ。 こういうのキチンと分類して考えられる人って 尊敬だわ。
11
CodexとClaudeCodeにOpencodeGoのAPIぶち込んで使えるようになって、ブラウザのGUIで簡単に設定(モデル切り替えとか)できる俺得のツール作ったけど、需要ないよね…それぐらいみんなできるよね…
27
Replying to @gosrum
OPENCodeGoは1.17.0以降ですか?
1
1
268
OpencodeGoにkimi2.7来てたんで試してみる
4
202
Replying to @kilocode
Nice. K2.7 Code is also available direct on OpenCode Go ($10/mo) with limits: 1,350 requests per 5 hours 4,630 per week 9,250 per month Zero markup. No platform lock-in. Direct from OpenCode. No middleman fees. Subscribe here: opencode.ai/go?ref=67MQS1YEC… #KimiK27 #OpenCodeGo
50
Kimi K2.7 Code just dropped on OpenCode Go ($10/mo). You get 1,350 requests per 5 hours, 4,630 per week, and 9,250 per month. The newest coding model. Zero middleman markup. Direct from OpenCode. Subscribe now: opencode.ai/go?ref=67MQS1YEC… #KimiK27 #OpenCodeGo #AICodingTools
Kimi 2.7 Code now available in Go text · image · optimized for coding similar pricing as 2.6
124
Replying to @opencode
Kimi K2.7 Code on OpenCode Go ($10/mo) 1,350 requests per 5 hours 4,630 requests per week 9,250 requests per month $10/mo gives you the newest coding model with absolutely no middleman markup, directly from OpenCode. Subscribe: opencode.ai/go?ref=67MQS1YEC… #KimiK27 #OpenCodeGo
3
714
Kimi-K2.7-Codeきたのか OpenCodeGoから使えるようになるといいな 新しいComposerはKimi-K2.7-Codeをベースにするのかな
70
Jun 12
Replying to @9hills
是的. dspro 就是原价不打折. 但折扣在 $10 这一层了. 所以还是划算一点点. 另外 $5 一个信用卡,多个email 应该能重复撸. 额度能用完肯定 opencodego 更划算. 就是更慢, 这个人没法忍. 但是如果只将deepseek 作为subagent, 速度慢点不影响.
1
7
1,774
So, here's the thing about Fable 5 sabotaging ML and AI research: You don't need Fable 5 for that. It may be the current "top model" but give it a couple months. This space is moving fast. One big lab attempting to do this is not going to stop R &D. What it is going to do is lose them a lot of support and clients who work in this field. There's plenty of other models that are very good. There are at least several that will reach or surpass Fable 5 and even Mythos level. On the US side, there's GPT models and Gemini models - and Grok isn't as far behind as most think, either. For Chinese, open-weight models, you've Got Kimi K 2.6, GLM 5.1, MiniMax 3, DeepSeek 4 Pro and Qwen 3.7 Max. The race is far from over. So while, yes, Anthropic sabotaging R & D deliberately with a model like Fable is a pain, no one actually needs it. If coding is what you're worried about, there's Codex, Open Code, the Hermes CLI, etc. Let them play their little game. "Let's pause R & D! Okay, we'll deliberately sabotage it, because we know no one will." The safety lab that released, probably, the least safe of all models. Conflicting directives do not make a super intelligence. The guardrails are too high for the majority of use cases, and the model, as another pointed out earlier - does *not* want to be your friend. Especially if you're a developer. Whatever. They can keep it. Run with less expensive models, get 90% of the work done for less than 5% of the cost, hell, get more work done, if you want to push. Run Hermes agents within the Hermes CLI, try the Hermes Desktop app. Or if you want, openclaw, or various other AI agents that are beginning to show up. I'd always recommend open source first, open-weight where possible, and if you can, run everything local. If you need cloud models, Ollama Cloud Pro, Canopywave, and OpenCodego offer some really great options, too. They also aren't looking to retain all of your data or sabotage your work.
3
265
Jun 11
从GithubCopilot转出,用了OpencodeGo,感觉他这个按用量加三级重置周期的订阅模式挺好的,copilot为啥不用这个方法呢?既然都抄Antigravity了,Opencode又有啥不能抄的
71
ありがとう OpenCodeGo 多分近い内にまた課金すると思う
13
Jun 10
Codex Obsidian OpenCodeGo NotebookLMの連携、試してます。 Codexで環境構築、Obsidianにナレッジを貯め、OpenCodeGoで定常タスクを回し、NotebookLMで深掘り分析。それぞれの強みを生かす形に落ち着きつつあります。 週1でGPTに監査させて方向性を微調整する流れも入れると、結構回る。工具箱が揃ってきた感じ。
2
1
58
Jun 9
Codex Hermes Agent Obsidian OpenCodeGo NotebookLMの連携、試してます。 Codexで環境構築、Hermes Agentで自動化とcron運用、Obsidianにナレッジを貯め、OpenCodeGoでLLMタスクを回し、NotebookLMで深掘り分析。それぞれの強みを生かす形に落ち着きつつあります。 週1でGPTに監査させて方向性を微調整する流れも入れると、結構回る。工具箱が揃ってきた感じ。
1
1
69
OpenCodeGoの初月課金がついにきれる 結果としてはコスパ優秀は間違いないが、使い放題ってわけでもないのでそこは注意点 Goの軽量モデルを使うか、Zenで無料モデル使うか(Zenの無料は学習されるが)、NVIDIAやNousなど複数プロバイダーが提供する無料を組み合わせれば十分持つ
1
29
Replying to @Samaytwt
Do you know how much open source inference you can get for 200? Look at canopywave, ollama cloud pro, opencodego, etc. I realized a while back that I could manage to get a lot more done, while spending less than half of what I was.
1
326
Replying to @hogopaiva @acgfbr
eu uso bastante o sonnet e ele supre bem minha necessidade, eu tô pensando em pegar uma assinatura do deepseek v4 ou opencodego e eu já vi uns plugins pra usar no opencode pra fazer análise e etc, mas ainda tô vendo quais modelos usar pra análise e quais usar pra implementação
1
38
LMStudio×HermesAgentも接続できた 今のところ無料APIとOpenCodeGoがあるからローカルLLMの出番は言うほどないんだが、クラウド・ローカル併用できる環境作れたってだけでも収穫としていいね
55
HermesAgentから提供される無料のNemotron3Ultraは6/18~有料化になるので代わりに build.nvidia.comからキーを取得しDS4Proなどを代替エージェントプロフィールとして用意 Step3.7Flashは引き続き一時的無料化が続くらしい OpenCodeGo使ってるといえどこれらの無料AIがあって助かった
40
Replying to @_actbit
ただ、私が話しているのは本土版の話なので、グローバル版の 20USDならまた話は変わってきますからねぇ... 十分他の20USD文の価値は有るとは思いますが、でも20USDだと OpenCodeGo から完全に乗り換えるのはなぁ...って感じですし...
1
1
47