★VALS INDEX
60.2%
+/- 31.9
Accuracy
⏱VALS INDEX
3.6s
Latency (median)
$VALS INDEX
—
Cost / Test (1P-est)
SCORE STATUSscored
TEST SETS13
RECORDED TRIALS673
CUSTOM PASS70%
Industry benchmarks
Accuracyattempts · responded
GPQA_DIAMOND
GPQA Diamond
59.4%
180 attempts · 180 responded
HUMANEVAL_PLUS
HumanEval+
95.0%
40 attempts · 40 responded
IFEVAL
IFEval
68.3%
180 attempts · 180 responded
MMLU_PRO
MMLU-Pro
18.0%
200 attempts · 200 responded
Custom 9-task suite
Mean /5n=trials
CUSTOM
json_strict
★
5.00
6 trials
CUSTOM
reasoning_multistep
★
5.00
6 trials
CUSTOM
agentic_prompt
★
4.82
11 trials
CUSTOM
summarize
★
4.82
11 trials
CUSTOM
code_gen_long
★
4.45
11 trials
CUSTOM
agentic_tool_use
★
4.33
3 trials
CUSTOM
refactor_existing_code
-
3.67
3 trials
CUSTOM
code_debug
-
2.91
11 trials
CUSTOM
creative_write
-
0.00
11 trials
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.