loop-benchmarking

Controlled experiments across agentic coding configurations. Same task, one variable, what actually works.
git clone https://git.shiptheloop.com/loop-benchmarking.git
Log | Files | Refs | README

commit e59ff443edb659c9d21d3fef8d708bd29176f827
parent 7a1efd6efd6f2649b557db27eb8af49fd4795d6e
Author: Brian Graham <brian@buildingbetterteams.de>
Date:   Wed,  8 Apr 2026 07:09:29 +0200

Add minimax-m2.7 and kimi-k2.5 via OpenRouter

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Diffstat:
Mdashboard/src/lib/colors.ts | 5+++++
Mgrid.yaml | 8+++++++-
Mharness/lib/compute_grid.py | 2++
3 files changed, 14 insertions(+), 1 deletion(-)

diff --git a/dashboard/src/lib/colors.ts b/dashboard/src/lib/colors.ts @@ -28,6 +28,11 @@ export const MODEL_COLORS: Record<string, string> = { // Google "gemma-4-26b": "hsl(60 50% 62%)", // olive + // MiniMax + "minimax-m2.7": "hsl(340 50% 65%)", // rose + // Moonshot + "kimi-k2.5": "hsl(170 40% 55%)", // teal + // Future slots "slot-7": "hsl(270 40% 68%)", // lavender "slot-8": "hsl(60 50% 62%)", // olive diff --git a/grid.yaml b/grid.yaml @@ -9,7 +9,7 @@ defaults: axes: model: - values: ["haiku-4.5", "sonnet-4.6", "opus-4.6", "glm-4.5-air", "glm-4.7", "glm-5.1", "qwen-3.6-plus", "gemma-4-26b"] + values: ["haiku-4.5", "sonnet-4.6", "opus-4.6", "glm-4.5-air", "glm-4.7", "glm-5.1", "qwen-3.6-plus", "gemma-4-26b", "minimax-m2.7", "kimi-k2.5"] effort: values: [high, max] prompt_style: @@ -72,6 +72,8 @@ providers: cli_model_map: "qwen-3.6-plus": "openrouter/qwen/qwen3.6-plus:free" "gemma-4-26b": "openrouter/google/gemma-4-26b-a4b-it" + "minimax-m2.7": "openrouter/minimax/minimax-m2.7" + "kimi-k2.5": "openrouter/moonshotai/kimi-k2.5" exclusions: # Haiku does not support extended thinking @@ -98,7 +100,11 @@ exclusions: # OpenRouter models only with openrouter - when: { provider: anthropic, model: "qwen-3.6-plus" } - when: { provider: anthropic, model: "gemma-4-26b" } + - when: { provider: anthropic, model: "minimax-m2.7" } + - when: { provider: anthropic, model: "kimi-k2.5" } - when: { provider: zai, model: "gemma-4-26b" } + - when: { provider: zai, model: "minimax-m2.7" } + - when: { provider: zai, model: "kimi-k2.5" } - when: { provider: openrouter, model: "haiku-4.5" } - when: { provider: openrouter, model: "sonnet-4.6" } - when: { provider: openrouter, model: "opus-4.6" } diff --git a/harness/lib/compute_grid.py b/harness/lib/compute_grid.py @@ -74,6 +74,8 @@ VALUE_ABBREV = { "opus-4.6": "opus46", "qwen-3.6-plus": "qwen36p", "gemma-4-26b": "gemma426b", + "minimax-m2.7": "mmx27", + "kimi-k2.5": "kimi25", "anthropic": "anth", "openrouter": "or", }

Impressum · Datenschutz