skynet · general
SKYNET

zai-coding/glm-5

ZAI GLM-5 (prior-gen)

⚖ COMPARE MODELS
VALS INDEX
14.1% +/- 13.7
Accuracy
VALS INDEX
31.5s
Latency (median)
$VALS INDEX
$0.3373
Cost / Test (1P-est)
SCORE STATUSscored
TEST SETS13
RECORDED TRIALS556
CUSTOM PASS61%

Industry benchmarks

Accuracyattempts · responded
GPQA_DIAMOND GPQA Diamond
2.0%
100 attempts · 11 responded
HUMANEVAL_PLUS HumanEval+
32.5%
40 attempts · 15 responded
IFEVAL IFEval
5.5%
100 attempts · 6 responded
MMLU_PRO MMLU-Pro
16.2%
260 attempts · 146 responded
Industry benchmark Bars = accuracy %, black cap = that score

Custom 9-task suite

Mean /5n=trials
CUSTOM json_strict
5.00
3 trials
CUSTOM reasoning_multistep
5.00
3 trials
CUSTOM code_gen_long
4.88
8 trials
CUSTOM summarize
4.25
8 trials
CUSTOM agentic_tool_use
4.00
5 trials
CUSTOM agentic_prompt
3.75
8 trials
CUSTOM code_debug -
2.25
8 trials
CUSTOM creative_write
1.88
8 trials
CUSTOM refactor_existing_code -
0.80
5 trials
Custom suite · scored 0–5 by consensus (m3 + Fable-5) ★ = task scored a 5 at least once · sorted by mean score

Cross-judge validation

m3 vs Fable-5 Δ

No judge bias detected on this model — m3 and Fable-5 agree across all tasks.

Per-task distribution

Score

All trials on this page were run through the harness. Industry suites use deterministic family-specific reference graders (no LLM judge). The custom suite is graded by consensus of minimax-m3 + claude-fable-5. See /industry/ for the full benchmark suite.