Make mercury-2.5-preview the default model, replacing retired gemini-2.5-flash-lite - #886
Merged
Conversation
Free OpenRouter model to try as a faster alternative to the current default. Single provider (Novita), 262k context, 32k max output. Not given a priority; select it manually when experimenting. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
General-purpose sibling of the sante model, pinned to the Novita endpoint. Cheaper than the current deepseek-v4-flash default. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Finance-focused free sibling of Ling 3.0 Flash, single provider Novita. Another candidate for the speed experiment. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Free open-weight MoE from Dots Studio, single provider AtlasCloud. Another candidate for the speed experiment. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
…free to baseline.json More candidates for the speed experiment. The paid nemotron entry is pinned to DeepInfra bf16. thinkingmachines/inkling-small:free was requested but omitted: OpenRouter returns 403, the model is restricted to allowlisted agentic harnesses. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Diffusion LLM from Inception, single provider. Candidate for the speed experiment. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Gemini 2.5 Flash Lite was retired. Full plan generation on the interim default (deepseek-v4-flash) took 3 to 4 hours versus about 15 minutes before. This branch trials a set of faster OpenRouter models and settles on Inception's Mercury 2.5 Preview, a diffusion LLM that generates tokens in parallel.
openrouter-mercury-2.5-previewis now priority 1.openrouter-solar-pro4is priority 2. Priorities 3 and 4 are unchanged.openrouter-ling-3.0-flash(pinned to Novita),openrouter-nemotron-3.5-lightning(pinned to DeepInfra bf16),openrouter-solar-pro4,openrouter-mercury-2.5-preview.thinkingmachines/inkling-small:freewas evaluated but not added: OpenRouter returns 403, it is restricted to allowlisted agentic harnesses.Verification
get_llm()with the real config; Mercury answered in about 1.4s with heavy hidden reasoning and no latency penalty.baseline.jsonparses; no keys removed relative to main.🤖 Generated with Claude Code