Skip to content

Refresh MiniMax model metadata - #1062

Open
octo-patch wants to merge 1 commit into
agentscope-ai:mainfrom
octo-patch:octo/20260719-refresh-minimax-model-metadata-recvoSrk2uBvuN
Open

Refresh MiniMax model metadata#1062
octo-patch wants to merge 1 commit into
agentscope-ai:mainfrom
octo-patch:octo/20260719-refresh-minimax-model-metadata-recvoSrk2uBvuN

Conversation

@octo-patch

Copy link
Copy Markdown
Contributor

Summary

  • add MiniMax-M3 to the built-in manager and worker model registries
  • refresh MiniMax-M2.7 and MiniMax-M2.7-highspeed context windows to 204,800 tokens
  • keep generated templates, model-switch scripts, installer detection, and controller defaults aligned
  • add focused controller assertions for MiniMax model metadata

Scope

AgentTeams does not maintain provider pricing, prompt-cache pricing, or direct MiniMax endpoint registries; its models route through the configured AgentTeams gateway. This PR updates the MiniMax fields represented by the repository’s existing model schema.

Validation

  • jq assertions for MiniMax-M3 and MiniMax-M2.7 in manager/configs/known-models.json
  • bash -n on the three changed runtime scripts and the Unix installer
  • git diff --check
  • repository-configured secret scan

go test ./internal/agentconfig could not start because the available Go toolchain rejects the repository’s go 1.25.0 directive before package compilation.

@github-actions

Copy link
Copy Markdown
Contributor

📊 CI Metrics Report

Summary

Metric Current Baseline Change
LLM Calls 80 97 -17 ↓ -17.5%
Input Tokens 2089934 2719498 -629564 ↓ -23.2%
Output Tokens 13805 16508 -2703 ↓ -16.4%
Total Tokens 2103739 2736006 -632267 ↓ -23.1%

By Role

Role Metric Current Baseline Change
🧠 Manager LLM Calls 68 72 -4 ↓ -5.6%
Input Tokens 1818388 2092724 -274336 ↓ -13.1%
Output Tokens 11015 10274 +741 ↑ +7.2%
Total Tokens 1829403 2102998 -273595 ↓ -13.0%
🔧 Workers LLM Calls 12 25 -13 ↓ -52.0%
Input Tokens 271546 626774 -355228 ↓ -56.7%
Output Tokens 2790 6234 -3444 ↓ -55.2%
Total Tokens 274336 633008 -358672 ↓ -56.7%

Per-Test Breakdown

Test Mgr Calls Wkr Calls Δ Calls Mgr In Wkr In Mgr Out Wkr Out Δ Tokens Trend
02-create-worker 6 0 +1 ↑ +20.0% 158254 0 991 0 +22257 ↑ +16.2% ⚠️ regressed
03-assign-task 10 6 +3 ↑ +23.1% 254211 130975 1459 942 +77407 ↑ +25.0% ⚠️ regressed
04-human-intervene 15 0 +1 ↑ +7.1% 332162 0 2322 0 -1953 ↓ -0.6% ⚠️ regressed
05-heartbeat 7 0 0 — 0% 192824 0 1127 0 -4253 ↓ -2.1% — unchanged
06-multi-worker 30 6 -22 ↓ -37.9% 880937 140571 5116 1848 -725725 ↓ -41.4% ✅ improved

Trends

1 test(s) improved (fewer LLM calls)
⚠️ 3 test(s) regressed (more LLM calls)


Generated by HiClaw CI on 2026-07-19 16:29:36 UTC


📦 Download debug logs & test artifacts

@github-actions

Copy link
Copy Markdown
Contributor

❌ Integration Tests Failed (llm-interaction-2 / mgr=copaw / wk=hermes)

Commit: dec83ca
Workflow run: #1700

Test Results
No test output captured.
Debug Log (tail)
No debug logs available.

📦 Download full debug logs & test artifacts

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant