Beam is now available in OpenCode
Beam is now available in OpenCode with a 1M-token context window, offered free for a limited time. The announcement matters to developers using OpenCode for coding workflows that need to process large codebases or long sessions in one context.
- Beam offers 1M-token context in OpenCode
- Beam is free for a limited time
Related
GLM 5.3 Flash is now free in OpenCode
GLM 5.3 Flash is now free in OpenCode
GLM 5.3 Flash is now free in OpenCode
Hit-or-Miss Alone, Solid in a Loop: 48 Hours with Reflection’s Beam
A Medium writeup of 48 hours using Reflection's Beam model finds it hit-or-miss on single attempts but reliable when run in loops that let it verify its own output. The reviewer describes the adjustable reasoning dial as the product's core feature and argues the advertised 100M-token context isn't the practical limit.
A Medium writeup of 48 hours using Reflection's Beam model finds it hit-or-miss on single attempts but reliable when run in loops that let it verify its own output. The reviewer describes the adjustable reasoning dial as the product's core feature and argues the advertised 100M-token context isn't the practical limit.
Meta Releases Muse Spark 1.3, matching Fable 5
Meta released Muse Spark 1.3 on 2026-09-02, targeting improved performance on agentic and coding tasks and real-world usability. The model rolled out in Muse Code and the Meta Model API, with the Contributor variant available on OpenCode Go and a max-reasoning mode planned after safety testing. A Reddit post reports pricing of $0.10 input / $0.20 output per million tokens, matching Fable 5.
Meta released Muse Spark 1.3 on 2026-09-02, targeting improved performance on agentic and coding tasks and real-world usability. The model rolled out in Muse Code and the Meta Model API, with the Contributor variant available on OpenCode Go and a max-reasoning mode planned after safety testing. A Reddit post reports pricing of $0.10 input / $0.20 output per million tokens, matching Fable 5.
Hy4 preview launches in OpenCode Go with 770B/49B parameters and 1M context
Hy4 preview is now available in OpenCode Go, announced via X on 2026-08-28. The model pairs 770B total with 49B active parameters and a 1M-token context window, positioned for coding agents. Availability in OpenCode Go gives agent developers a new large-context coding model option.
Hy4 preview is now available in OpenCode Go, announced via X on 2026-08-28. The model pairs 770B total with 49B active parameters and a 1M-token context window, positioned for coding agents. Availability in OpenCode Go gives agent developers a new large-context coding model option.
LongCat-2.5-Preview is now free on OpenCode for two weeks
LongCat-2.5-Preview is available for free on OpenCode for a two-week window. The offer highlights 1M-token context, multi-modal input, and zero data retention. It lowers the barrier for developers to trial the model's long-context and multimodal capabilities without data-retention concerns.
LongCat-2.5-Preview is available for free on OpenCode for a two-week window. The offer highlights 1M-token context, multi-modal input, and zero data retention. It lowers the barrier for developers to trial the model's long-context and multimodal capabilities without data-retention concerns.
