Loading model…
z-ai/glm-5.270.8%, pass rate 50.4% (231/458)3.2s, output speed 27.3 tok/s$1.16, prompt tokens 29352, completion tokens 25327927/30 passed in cost category with minimal spend ($0.03)52/60 passed, output speed 27.3 tok/s across 458 tests54/60 passed (247/261 score)7/60 passed (342/541 score), the lowest category pass rate7/60 passed (324/541 score)17/60 passed (328/512 score), suggesting vulnerability-handling issues2/8 passed, the smallest sample but worst ratioglm-5.2 at 70.8% / 50.4% trails gpt-5.5 (77.9% / 61.9%), gpt-5.3-codex (77.7% / 61.2%), and qwen3.7-max (75.3% / 54.9%), with a notable ~11-point pass-rate gap to first place.Overall score % from merged run_models rows (chronological). Only runs that include this model appear as points.
Score % vs pass rate % per category. With 0/1 scorers, both usually line up; with proportional tests, score % reflects partial credit while pass rate counts tests that clear the fixture threshold.
Total estimated spend per scope for this model (bars, left axis) and mean spend per merged result row (line, right axis: total ÷ tests).
Pass rate % per difficulty level — complements the score % view above.
Normalized 0–100 within this model: TTFT (shorter → higher spoke) and decode tok/s (higher → higher spoke). Values come from streamed BLXBench runs merged into overall_ranking.json.
Pass rate % per category for this model (distinct from score %, which reflects partial credit).