skynet · fast
SKYNET

zai-coding/glm-5-turbo

ZAI GLM 5 turbo (fast tier, untested)

⚖ COMPARE MODELS
VALS INDEX
40.2% +/- 33.1
Accuracy
VALS INDEX
60.5s
Latency (median)
$VALS INDEX
$0.2823
Cost / Test (1P-est)
SCORE STATUSscored
TEST SETS12
RECORDED TRIALS165
CUSTOM PASS60%
DEEP VALIDATION · 2026-08-20

Dual-judge family campaign

FULL REPORT →
CONSENSUS MEAN4.167 / 5
BOTH-JUDGE PASS75.6%
CROSS-JUDGED45 / 45
MEDIAN WALL37.40s
GPQA_DIAMONDserialized verification sample
80.0%
23/25 responses
IFEVALserialized verification sample
80.0%
25/25 responses
MMLU_PROserialized verification sample
48.0%
24/25 responses

Industry benchmarks

Accuracyattempts · responded
GPQA_DIAMOND GPQA Diamond
3.3%
30 attempts · 1 responded
IFEVAL IFEval
67.2%
30 attempts · 27 responded
MMLU_PRO MMLU-Pro
50.0%
60 attempts · 60 responded
Industry benchmark Bars = accuracy %, black cap = that score

Custom 9-task suite

Mean /5n=trials
CUSTOM agentic_tool_use
5.00
5 trials
CUSTOM code_gen_long
5.00
5 trials
CUSTOM json_strict
5.00
5 trials
CUSTOM reasoning_multistep
5.00
5 trials
CUSTOM summarize
4.20
5 trials
CUSTOM creative_write
2.80
5 trials
CUSTOM refactor_existing_code -
2.20
5 trials
CUSTOM code_debug -
2.00
5 trials
CUSTOM agentic_prompt -
1.20
5 trials
Custom suite · scored 0–5 by consensus (m3 + Fable-5) ★ = task scored a 5 at least once · sorted by mean score

Per-task distribution

Score

The accumulated custom suite below uses minimax-m3 + claude-fable-5 consensus. The dedicated 2026-08-20 campaign above is supplemental and uses minimax-m3 + a blind grok-4.6 re-judge on every successful response. Industry suites use deterministic family-specific graders. See methodology and the campaign report.