★VALS INDEX
75.7%
+/- 17.5
Accuracy
⏱VALS INDEX
7.7s
Latency (median)
$VALS INDEX
$1.8118
Cost / Test (1P-est)
SCORE STATUSscored
TEST SETS11
RECORDED TRIALS265
CUSTOM PASS64%
Industry benchmarks
Accuracyattempts · responded
GPQA_DIAMOND
GPQA Diamond
63.3%
30 attempts · 30 responded
HUMANEVAL_PLUS
HumanEval+
88.0%
50 attempts · 50 responded
Custom 9-task suite
Mean /5n=trials
CUSTOM
code_gen_long
★
5.00
20 trials
CUSTOM
json_strict
★
5.00
20 trials
CUSTOM
reasoning_multistep
★
5.00
20 trials
CUSTOM
agentic_tool_use
★
4.70
20 trials
CUSTOM
summarize
★
3.70
20 trials
CUSTOM
code_debug
★
3.55
20 trials
CUSTOM
agentic_prompt
★
3.10
20 trials
CUSTOM
refactor_existing_code
-
2.60
25 trials
CUSTOM
creative_write
★
1.55
20 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.