Meta's Muse Wonder tops DeepSWE coding agent leaderboard at 78.8%, ahead of Claude Opus and GPT-6
AccBalanced · x · 2026-10-20
OpenRouter announced Meta's Muse Wonder is live, as the DeepSWE benchmark (113 original long-horizon engineering tasks across 91 repos) published its v1.1 leaderboard: Muse Wonder (max) takes #1 at 78.8%, ahead of Claude Opus 5.5 (max) at 74.2%, GPT-6 Astra (max) at 74.1%, and Claude Fable 5.1 (max) at 67.4%.
Full standings: gpt-6-astra 74% (avg cost $4.43), gemini-3.8-flash 74% ($2.36), claude-opus-5 74% ($11.84), gpt-5.6-sol 73%, claude-fable-5 70%, glm-5.3 69%, kimi-k3 69%, grok-4.6 67%, deepseek-v4-pro 63%. Best value picks: gpt-5.6-luna (67% at $0.61) and glm-5.3-flash (63% at $0.24); claude-sonnet-5 scores just 54% at a steep $26.40 per task.
More from Models
- REAX inference engine runs Perplexity Decider up to 6x faster, free for individuals — airesearch12 · 2026-10-20
- SKY-001 debuts as first 'Engineering World Model,' claiming 125x speedup over GPT-6 on CAD — vasuman · 2026-10-20
- Smart glasses test: Octen search at 60ms, model generation is the bottleneck; Mercury 3 wins — Scobleizer · 2026-10-20
- BitNet CPU inference is silently wrong: perplexity 6x reference, one-line fix works — matheus-c-feitosa · 2026-10-20
- Space Bunny stealth model revealed as MiniMax M3.1 Flash: 428B-A23B, weights Thursday — BTA_Labs · 2026-10-20
- Clairic is live on Workers AI; open-weights release planned once the architecture matures — michellechen · 2026-10-20