docs(which-model): add muse-spark-1.3 and deepseek-v4.1-flash lanes - #40
Merged
Merged
Conversation
Verified 2026-09-12 against the live OpenRouter models API and the running opencode catalog: openrouter/meta/muse-spark-1.3 ($1.25/$4.25, 1M ctx, $0.15 cached input, AA index 62) joins frontier synthesis and the long-context heuristic; openrouter/deepseek/deepseek-v4.1-flash ($0.15/$0.60, 1M ctx, $0.003 cached) leads the cheap-code lane. The frontier-reservation rule gains the Muse Spark caveats: mandatory reasoning bills hidden CoT at the output rate, and the Contributor SKU trains on prompts, so it must never carry customer data.
cheshirecode
added a commit
that referenced
this pull request
Sep 13, 2026
Companion to PR #40's routing change: deepseek-v4.1-flash ($0.15/$0.60 per MTok, 1M ctx, $0.003 cached input) undercuts qwen3.7-plus ($0.32/$1.28) on every axis. Primary model stays glm-5.3-flash. Availability verified against opencode models and the live OpenRouter API 2026-09-12.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Adds two live-verified models to the opencode lane table in
skills/which-model/references/routing.md:openrouter/meta/muse-spark-1.3→ frontier-synthesis lane + long-context heuristic (1M ctx, $1.25/$4.25, $0.15 cached input, AA Intelligence Index 62, verified against the live OpenRouter models API 2026-09-12)openrouter/deepseek/deepseek-v4.1-flash→ cheap-code lane, first position ($0.15/$0.60, 1M ctx, $0.003 cached)The frontier-reservation guardrail gains Muse Spark's two operational caveats: mandatory reasoning bills hidden CoT at the output rate, and the Contributor SKU (
-contributor-free, $0.10/$0.20) trains on prompts — never route customer data or proprietary source there.Companion config change (outside the repo)
~/.config/opencode/opencode.jsonc:small_model→openrouter/deepseek/deepseek-v4.1-flash(2× cheaper thanqwen3.7-plusit replaces). Primary staysglm-5.3-flash. Muse Spark deliberately not a default — mid-tier pricing, escalation-lane only.Verification
opencode modelsand the OpenRouter API response (2026-09-12)python3 -m unittest discover -s skills/which-model/tests— OK