Skip to content

Add models#32

Merged
Yuki-Imajuku merged 5 commits into
mainfrom
feat/add-models
Apr 10, 2026
Merged

Add models#32
Yuki-Imajuku merged 5 commits into
mainfrom
feat/add-models

Conversation

@Yuki-Imajuku

Copy link
Copy Markdown
Collaborator

Added models:

  • Claude 4.6
  • Qwen3.5, 3.6
  • Nemotoron3
  • GLM-5.1
  • Gemma 4

Small changes:

  • Set default timeout to 1h
  • update dependencies

Copilot AI review requested due to automatic review settings April 10, 2026 23:17

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

This PR expands the benchmark’s supported LLM set by adding/updating llm_configs/ entries and extending the internal fallback pricing table used for cost estimation, along with a longer default inference timeout and minor dependency/tooling updates.

Changes:

  • Added/updated model configs for Claude 4.6, Qwen 3.5/3.6, Nemotron 3, GLM-5.1, and Gemma 4 (plus a GPT-5.4 “medium” config).
  • Extended calc_cost fallback pricing map with new model price entries (and adjusted Claude 4.6 pricing structure).
  • Increased scaffold default timeout from 20m to 1h; updated dev tooling deps and added a uv setting.

Reviewed changes

Copilot reviewed 12 out of 13 changed files in this pull request and generated 4 comments.

Show a summary per file
File Description
src/ale_bench_eval/scaffolds.py Increase default generation timeout used by repeated-sampling/self-refine flows.
src/ale_bench_eval/calc_cost.py Add new fallback prices (Gemma 4, GLM-5.1, Nemotron 3, Qwen 3.5/3.6) and adjust Claude 4.6 pricing entries.
pyproject.toml Bump dev tool versions and add [tool.uv] exclude-newer configuration.
llm_configs/qwen3.6-plus.json Update config to target Qwen 3.6 Plus on OpenRouter with Alibaba routing.
llm_configs/qwen3.5-35b-a3b.json New OpenRouter config for Qwen 3.5 35B A3B.
llm_configs/nemotron-3-super.json New OpenRouter config for Nemotron 3 Super (free).
llm_configs/gpt-5.4-medium.json New OpenAI config for GPT-5.4 medium reasoning profile.
llm_configs/glm-5.1.json Update config to target GLM-5.1 on OpenRouter.
llm_configs/gemma-4-26b-a4b-it.json Update config to target Gemma 4 26B on OpenRouter with Parasail routing.
llm_configs/claude-4.6-sonnet-medium.json Expand provider routing order and adjust verbosity setting.
llm_configs/claude-4.6-opus-no-thinking.json New Anthropic-direct config for Claude Opus 4.6 with thinking disabled.
llm_configs/claude-4.6-opus-medium.json New Anthropic-direct config for Claude Opus 4.6 with adaptive thinking.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

@SakanaAI SakanaAI deleted a comment from Copilot AI Apr 10, 2026
@SakanaAI SakanaAI deleted a comment from Copilot AI Apr 10, 2026
@SakanaAI SakanaAI deleted a comment from Copilot AI Apr 10, 2026
@SakanaAI SakanaAI deleted a comment from Copilot AI Apr 10, 2026
@Yuki-Imajuku Yuki-Imajuku merged commit 1372a02 into main Apr 10, 2026
14 checks passed
@Yuki-Imajuku Yuki-Imajuku deleted the feat/add-models branch April 10, 2026 23:23
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.

2 participants