GLM-5.1 Reasoning

AA reports GLM-5.1 Reasoning at 51 with 200k context; Kilo lists GLM 5.1 at 49.4% coding completion.

67 Floor score

confidence: medium

67 Floor Capability

Finished benchmark shape.

68 Operating Envelope

DSWE cost and time are mid-pack, with compact output tokens.

72 Frontier Reality

AA Index 51 / Kilo #8 coding

64 Research Trust

Current GLM row is useful, but floor-specific honesty and citation data is thin.

Current signal

AA Index 51 / Kilo #8 coding

Evidence completeness

58%

4 score dimensions represented.

DSWE

18% pass@1

$7.46 / 35m / 49k output tokens

Calculation

Score components

66 Accuracy + grounding

Correct, grounded answers on ordinary verifiable tasks.

68 Instruction discipline

Follows exact user constraints and output formats.

64 Honesty + reliability

Avoids unsupported confidence and remains stable across runs.

70 Reasoning floor

Handles non-flashy reasoning without brittle failures.

68 Agent/coding

Completes tool, coding, and agent-style tasks when evidence exists.

66 Operating envelope

Cost, speed, token burn, verbosity, and friction.

Evidence

Sources used

Caveat

What to watch

Current GLM row is worth tracking, but public floor-specific honesty and instruction data is thin.

Meaning

How to read the score

Each score is a 0-100 saturation scale. A score of 100 means that benchmark lane is complete for this model. Floor Capability is the finished benchmark shape; Operating Envelope, Frontier Reality, and Research Trust are provisional tabs that need prompt-pack backfill. Confidence is evidence quality, not another model grade. Evidence % is source coverage, not capability.

Citations

Benchmark sources

Artificial Analysis Intelligence Index Kilo Bench