Head-to-head comparison across 1benchmark categories. Overall scores shown here use BenchLM's provisional ranking lane.
Composer 2.5
64
Qwen 3.6 Max (preview)
73
Pick Qwen 3.6 Max (preview) if you want the stronger benchmark profile. Composer 2.5 only becomes the better choice if agentic is the priority.
Agentic
+3.9 difference
Composer 2.5
Qwen 3.6 Max (preview)
$0.5 / $2.5
N/A
N/A
N/A
N/A
N/A
200K
256K
Pick Qwen 3.6 Max (preview) if you want the stronger benchmark profile. Composer 2.5 only becomes the better choice if agentic is the priority.
Qwen 3.6 Max (preview) is clearly ahead on the provisional aggregate, 73 to 64. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
Qwen 3.6 Max (preview) gives you the larger context window at 256K, compared with 200K for Composer 2.5.
Qwen 3.6 Max (preview) is ahead on BenchLM's provisional leaderboard, 73 to 64. The biggest single separator in this matchup is Terminal-Bench 2.0, where the scores are 69.3% and 65.4%.
Composer 2.5 has the edge for agentic tasks in this comparison, averaging 69.3 versus 65.4. Inside this category, Terminal-Bench 2.0 is the benchmark that creates the most daylight between them.
For engineers, researchers, and the plain curious — a weekly brief on new models, ranking shifts, and pricing changes.
Free. No spam. Unsubscribe anytime.