OpenCode Launches Phase 4 of Cheepseek: $10 Subscription Gets $100 of DeepSeek V4.1…
OpenCode has launched Phase 4 of Operation Cheepseek. Users of the $10-per-month OpenCode Go subscription can receive $100 worth of DeepSeek V4.1 Flash model credits, a step-by-step increase from the $30 offered in Phase 1. DeepSeek V4.1 Flash is a native multimodal MoE model released by DeepSeek in September, with a total parameter count of 552B and support for a context window of up to about 1 million tokens. The credits are converted based on model invocation billing, and the actual usable amount is still subject to request frequency, context length, and platform rate limits.
Body · AI translation
OpenCode Launches Phase 4 of Cheepseek: $10 Subscription Gets $100 of DeepSeek V4.1 Flash Credits
English text is being prepared.
English text is being prepared.
English text is being prepared.
The plan is mainly aimed at OpenCode Go subscribers paying $10 per month (approximately 66.7 yuan at the current exchange rate), allowing them to obtain large amounts of DeepSeek V4.1 Flash model usage at a low price:
Phase 1: Increase DeepSeek V4.1 Flash usage, with credits worth $30 (approximately 200.1 yuan at the current exchange rate)
Phase 2: Make the originally time-limited increased credits permanent, with credits worth $60 (approximately 400.2 yuan at the current exchange rate)
Phase 3: In celebration of OpenCode reaching 100,000 GitHub Stars, remove peak and off-peak restrictions, with credits worth $80 (approximately 533.6 yuan at the current exchange rate)
Phase 4: Once again increase the in-subscription model usage, creating leverage of "a $10 (approximately 66.7 yuan at the current exchange rate) subscription for $100 (approximately 667 yuan at the current exchange rate) of model credits," with credits worth $100 (approximately 667 yuan at the current exchange rate)
DeepSeek V4.1 Flash is a native multimodal mixture-of-experts model released by DeepSeek in September, with a total parameter count of 552B, supporting text and image input, function calling, agentic workflows, and a context window of up to about 1 million tokens. It activates about 8B parameters per token on the input side and about 16B parameters on the output side, thus offering lower inference costs for long-context and agentic tasks.
For ordinary developers and users of AI coding tools, the most direct impact of the credit increase is that, under the same OpenCode subscription, they can complete more code generation, code review, document understanding, multi-turn tool calling, and long-document analysis tasks.
However, the $100 (approximately 667 yuan at the current exchange rate) credit is not equivalent to a cash balance; it is usage value converted according to model call billing. The actual usable amount will still be affected by request frequency, context length, input and output scale, and platform rate limits.