From 5c1f18bcebb6966ea00f1460a63898194c9324df Mon Sep 17 00:00:00 2001 From: Vlad Ionescu Date: Sat, 6 Jun 2026 23:28:26 +0300 Subject: [PATCH] opencode: Model updates (#58743) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit OpenCode models updates (delayed since I was super-busy the past 2 weeks and nobody else opened up a PR): **Free**: - added MiniMax M3 Free as per [models.dev](https://github.com/anomalyco/models.dev/blob/f753ddacdbab785c87f4fceb3187d6ceb9790966/providers/opencode/models/minimax-m3-free.toml) and [docs](https://opencode.ai/docs/zen/). Seeing how MiniMax M2.5 was a long-term Free model, I am also assuming the same will be true for M3 🤷 Tested and confirmed working with a simple "_rename this variable for me. add a function. delete the function_" test - added Nemotron 3 Ultra Free as per [models.dev](https://github.com/anomalyco/models.dev/blob/f753ddacdbab785c87f4fceb3187d6ceb9790966/providers/opencode/models/nemotron-3-ultra-free.toml) and [docs](https://opencode.ai/docs/zen/), with 3 reasoning levels discovered manually from the OpenCode CLI. Tested and confirmed working with a simple "_rename this variable for me. add a function. delete the function_" test, both on `low` and `high` reasoning levels - removed Nemotron 3 Super Free as per [models.dev](https://github.com/anomalyco/models.dev/commit/72fc81a4daf22bccd14b383971e8bc92bbe2e64e) **Go**: - added MiniMax M3 as per [models.dev](https://github.com/anomalyco/models.dev/blob/f753ddacdbab785c87f4fceb3187d6ceb9790966/providers/opencode-go/models/minimax-m3.toml). Tested and confirmed working with a simple "_rename this variable for me. add a function. delete the function_" test - added Qwen 3.7 Plus as per [models.dev](https://github.com/anomalyco/models.dev/blob/f753ddacdbab785c87f4fceb3187d6ceb9790966/providers/opencode-go/models/qwen3.7-plus.toml). Tested and confirmed working with a simple "_rename this variable for me. add a function. delete the function_" test - updated token counts for Qwen 3.6 Plus as per https://github.com/anomalyco/models.dev/commit/350704d76271d3a166d85adc646006ec10bdc42c - removed Qwen 3.5 Plus as per [models.dev](https://github.com/anomalyco/models.dev/commit/c7af5ba9bee373a5a90d52b24d81276d778b48e2) (but not from Zen where [it is still valid](https://github.com/anomalyco/models.dev/blob/e524b01b9a8f593838f9d36469f0c5087ae2cf34/providers/opencode/models/qwen3.5-plus.toml)) **Zen**: - added DeepSeek V4 Flash as per [models.dev](https://github.com/anomalyco/models.dev/blob/f753ddacdbab785c87f4fceb3187d6ceb9790966/providers/opencode/models/deepseek-v4-flash.toml) and [docs](https://opencode.ai/docs/zen/). Not tested as I don't have a Zen subscription and I am stubbornly refusing to get one - added MiniMax M2.7 which was mistakenly marked as Go-only. Not tested as I don't have a Zen subscription and I am stubbornly refusing to get one ------ Self-Review Checklist: - [x] I've reviewed my own diff for quality, security, and reliability - [x] Unsafe blocks (if any) have justifying comments - [x] The content is consistent with the [UI/UX checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist) - [ ] Tests cover the new/changed behavior - [x] Performance impact has been considered and is acceptable Release Notes: - Agent: Updated OpenCode models (added MiniMax M3, Qwen 3.7 Plus, DeepSeek V4 Flash, MiniMax M3 Free, and Nemotron 3 Ultra Free; removed Nemotron 3 Super Free, Qwen 3.5 Plus; updated MiniMax M2.7, MiniMax M2.5, Qwen 3.6 Plus) --- crates/opencode/src/opencode.rs | 79 ++++++++++++++++++++++++--------- 1 file changed, 57 insertions(+), 22 deletions(-) diff --git a/crates/opencode/src/opencode.rs b/crates/opencode/src/opencode.rs index 91a89840bb..75e40f931a 100644 --- a/crates/opencode/src/opencode.rs +++ b/crates/opencode/src/opencode.rs @@ -141,18 +141,24 @@ pub enum Model { KimiK2_6, #[serde(rename = "minimax-m2.7")] MiniMaxM2_7, + #[serde(rename = "minimax-m3")] + MiniMaxM3, + #[serde(rename = "minimax-m3-free")] + MiniMaxM3Free, #[serde(rename = "mimo-v2.5-pro")] MimoV2_5Pro, #[serde(rename = "mimo-v2.5")] MimoV2_5, #[serde(rename = "big-pickle")] BigPickle, - #[serde(rename = "nemotron-3-super-free")] - Nemotron3SuperFree, + #[serde(rename = "nemotron-3-ultra-free")] + Nemotron3UltraFree, #[serde(rename = "qwen3.5-plus")] Qwen3_5Plus, #[serde(rename = "qwen3.6-plus")] Qwen3_6Plus, + #[serde(rename = "qwen3.7-plus")] + Qwen3_7Plus, #[serde(rename = "qwen3.7-max")] Qwen3_7Max, @@ -188,7 +194,7 @@ impl Model { } pub fn default_free_fast() -> Self { - Self::Nemotron3SuperFree + Self::Nemotron3UltraFree } pub fn available_subscriptions(&self) -> &'static [OpenCodeSubscription] { @@ -199,19 +205,22 @@ impl Model { | Self::KimiK2_6 | Self::KimiK2_5 | Self::MiniMaxM2_5 - | Self::Qwen3_5Plus + | Self::MiniMaxM2_7 + | Self::DeepSeekV4Flash | Self::Qwen3_6Plus => &[OpenCodeSubscription::Zen, OpenCodeSubscription::Go], // Go-only models - Self::MiniMaxM2_7 - | Self::MimoV2_5Pro + Self::MimoV2_5Pro | Self::MimoV2_5 | Self::DeepSeekV4Pro - | Self::DeepSeekV4Flash - | Self::Qwen3_7Max => &[OpenCodeSubscription::Go], + | Self::Qwen3_7Plus + | Self::Qwen3_7Max + | Self::MiniMaxM3 => &[OpenCodeSubscription::Go], // Free models - Self::Nemotron3SuperFree | Self::BigPickle => &[OpenCodeSubscription::Free], + Self::Nemotron3UltraFree | Self::BigPickle | Self::MiniMaxM3Free => { + &[OpenCodeSubscription::Free] + } // Custom models get their subscription from settings, not from here Self::Custom { .. } => &[], @@ -264,13 +273,16 @@ impl Model { Self::KimiK2_5 => "kimi-k2.5", Self::KimiK2_6 => "kimi-k2.6", Self::MiniMaxM2_7 => "minimax-m2.7", + Self::MiniMaxM3 => "minimax-m3", Self::MimoV2_5Pro => "mimo-v2.5-pro", Self::MimoV2_5 => "mimo-v2.5", Self::Qwen3_5Plus => "qwen3.5-plus", Self::Qwen3_6Plus => "qwen3.6-plus", + Self::Qwen3_7Plus => "qwen3.7-plus", Self::Qwen3_7Max => "qwen3.7-max", Self::BigPickle => "big-pickle", - Self::Nemotron3SuperFree => "nemotron-3-super-free", + Self::Nemotron3UltraFree => "nemotron-3-ultra-free", + Self::MiniMaxM3Free => "minimax-m3-free", Self::Custom { name, .. } => name, } @@ -319,13 +331,16 @@ impl Model { Self::KimiK2_5 => "Kimi K2.5", Self::KimiK2_6 => "Kimi K2.6", Self::MiniMaxM2_7 => "MiniMax M2.7", + Self::MiniMaxM3 => "MiniMax M3", Self::MimoV2_5Pro => "MiMo V2.5 Pro", Self::MimoV2_5 => "MiMo V2.5", Self::Qwen3_5Plus => "Qwen3.5 Plus", Self::Qwen3_6Plus => "Qwen3.6 Plus", + Self::Qwen3_7Plus => "Qwen3.7 Plus", Self::Qwen3_7Max => "Qwen3.7 Max", Self::BigPickle => "Big Pickle", - Self::Nemotron3SuperFree => "Nemotron 3 Super Free", + Self::Nemotron3UltraFree => "Nemotron 3 Ultra Free", + Self::MiniMaxM3Free => "MiniMax M3 Free", Self::Custom { name, display_name, .. @@ -375,7 +390,9 @@ impl Model { Self::Gemini3_1Pro | Self::Gemini3Flash | Self::Gemini3_5Flash => ApiProtocol::Google, - Self::Qwen3_7Max => ApiProtocol::Anthropic, + Self::Qwen3_7Max | Self::Qwen3_7Plus => ApiProtocol::Anthropic, + + Self::MiniMaxM3 | Self::MiniMaxM3Free => ApiProtocol::Anthropic, Self::Glm5 | Self::Glm5_1 @@ -389,7 +406,7 @@ impl Model { | Self::DeepSeekV4Pro | Self::DeepSeekV4Flash | Self::BigPickle - | Self::Nemotron3SuperFree => ApiProtocol::OpenAiChat, + | Self::Nemotron3UltraFree => ApiProtocol::OpenAiChat, Self::Custom { protocol, .. } => *protocol, } @@ -405,7 +422,9 @@ impl Model { | Self::MimoV2_5Pro | Self::Glm5 | Self::Glm5_1 - | Self::Nemotron3SuperFree + | Self::MiniMaxM2_5 + | Self::MiniMaxM2_7 + | Self::Nemotron3UltraFree | Self::BigPickle => true, Self::Custom { @@ -446,6 +465,8 @@ impl Model { // OpenAI-compatible models Self::MiniMaxM2_7 => 204_800, + Self::MiniMaxM3 => 512_000, + Self::MiniMaxM3Free => 200_000, Self::MiniMaxM2_5 => 204_800, Self::Glm5 | Self::Glm5_1 => { if subscription == OpenCodeSubscription::Go { @@ -458,10 +479,17 @@ impl Model { Self::GrokBuild0_1 => 256_000, Self::MimoV2_5Pro => 1_048_576, Self::MimoV2_5 => 1_000_000, - Self::Qwen3_5Plus | Self::Qwen3_6Plus => 262_144, - Self::Qwen3_7Max => 1_000_000, + Self::Qwen3_5Plus => 262_144, + Self::Qwen3_6Plus => { + if subscription == OpenCodeSubscription::Go { + 1_000_000 + } else { + 262_144 + } + } + Self::Qwen3_7Max | Self::Qwen3_7Plus => 1_000_000, Self::BigPickle => 200_000, - Self::Nemotron3SuperFree => 204_800, + Self::Nemotron3UltraFree => 1_000_000, Self::DeepSeekV4Pro | Self::DeepSeekV4Flash => 1_000_000, Self::Custom { max_tokens, .. } => *max_tokens, @@ -503,6 +531,8 @@ impl Model { // OpenAI-compatible models Self::MiniMaxM2_7 => Some(131_072), + Self::MiniMaxM3 => Some(131_072), + Self::MiniMaxM3Free => Some(32_000), Self::MiniMaxM2_5 => { if subscription == OpenCodeSubscription::Go { Some(65_536) @@ -520,9 +550,11 @@ impl Model { Self::BigPickle => Some(32_000), Self::KimiK2_6 | Self::KimiK2_5 => Some(65_536), Self::GrokBuild0_1 => Some(256_000), - Self::Qwen3_7Max | Self::Qwen3_6Plus | Self::Qwen3_5Plus => Some(65_536), + Self::Qwen3_7Max | Self::Qwen3_7Plus | Self::Qwen3_6Plus | Self::Qwen3_5Plus => { + Some(65_536) + } Self::DeepSeekV4Pro | Self::DeepSeekV4Flash => Some(384_000), - Self::Nemotron3SuperFree => Some(128_000), + Self::Nemotron3UltraFree => Some(128_000), Self::MimoV2_5Pro | Self::MimoV2_5 => Some(128_000), Self::Custom { @@ -578,7 +610,10 @@ impl Model { | Self::GrokBuild0_1 | Self::MimoV2_5 | Self::Qwen3_5Plus - | Self::Qwen3_6Plus => true, + | Self::Qwen3_6Plus + | Self::Qwen3_7Plus + | Self::MiniMaxM3 + | Self::MiniMaxM3Free => true, // OpenAI-compatible models without image support Self::MiniMaxM2_5 @@ -590,7 +625,7 @@ impl Model { | Self::DeepSeekV4Flash | Self::Qwen3_7Max | Self::BigPickle - | Self::Nemotron3SuperFree => false, + | Self::Nemotron3UltraFree => false, Self::Custom { protocol, .. } => matches!( protocol, @@ -611,7 +646,7 @@ impl Model { ReasoningEffort::XHigh, ]), - Self::MimoV2_5Pro | Self::MimoV2_5 => Some(vec![ + Self::Nemotron3UltraFree | Self::MimoV2_5Pro | Self::MimoV2_5 => Some(vec![ ReasoningEffort::Low, ReasoningEffort::Medium, ReasoningEffort::High,