In a 12-task Editorial Craft evaluation, Z.ai’s GLM 5.3 Flash ranked first with a 0.94 score and an estimated cost of $0.0002 per task. Step-3.7-flash followed at 0.92, while Qwen3.8 Flash, Claude Opus 5 (Fast), Gemini 3.7 Flash, and DeepSeek-V4-Flash-0731 rounded out the leaderboard.
By RuntimeWire Staff · Published
Leaderboard
| Rank | Model | Mean score | Est. cost/task |
|---|---|---|---|
| 1 | Z.ai: GLM 5.3 Flash | 0.94 | $0.0002 |
| 2 | step-3.7-flash | 0.92 | — |
| 3 | Qwen: Qwen3.8 Flash | 0.90 | $0.0003 |
| 4 | Claude Opus 5 (Fast) | 0.88 | $0.0334 |
| 5 | Google: Gemini 3.7 Flash | 0.87 | $0.0011 |
| 6 | DeepSeek-V4-Flash-0731 | 0.72 | — |
How we scored it
Every model answered the same 12-task battery from Editorial Craft, one task at a time, with no tools and no retries on content.
2 tasks carry ground truth (exact, numeric, multiple-choice, or the benchmark's official matcher) and were scored deterministically against the reference answer — correct is 1, incorrect is 0.
10 open-ended tasks were graded 0–10 by gpt-5.4 against a fixed rubric — blind to which model wrote the answer — and normalized to the same 0–1 scale.
Reasoning effort was pinned to max for every model whose lane exposes a control (OpenAI-style reasoning_effort
, Anthropic extended thinking, Gemini thinking config). Models marked vendor default on the interactive board expose no such control and ran as shipped.
A model's mean score averages its graded tasks; a generation failure counts as 0. Cost per task is estimated from each model's published per-token pricing ("—" where pricing isn't public), so treat it as directional, not billing-exact.
Explore every prompt, answer, and per-task grade in the interactive leaderboard.