Gemini 3.5 Flash

AA changelog lists Gemini 3.5 Flash at 55; Kilo lists it #4 coding at 64.7% completion. Public coverage emphasizes speed and broad app availability.

73 Floor score

confidence: medium

73 Floor Capability

Finished benchmark shape.

78 Operating Envelope

Fast DSWE time, but output token volume is very high.

79 Frontier Reality

AA Index 55 / Kilo #4 coding

68 Research Trust

Useful current Google row; needs internal hallucination and instruction-following runs.

Current signal

AA Index 55 / Kilo #4 coding

Evidence completeness

68%

5 score dimensions represented.

DSWE

28% pass@1

$7.42 / 17m / 189k output tokens

Calculation

Score components

70 Accuracy + grounding

Correct, grounded answers on ordinary verifiable tasks.

72 Instruction discipline

Follows exact user constraints and output formats.

68 Honesty + reliability

Avoids unsupported confidence and remains stable across runs.

74 Reasoning floor

Handles non-flashy reasoning without brittle failures.

78 Agent/coding

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

76 Operating envelope

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

Evidence

Sources used

Caveat

What to watch

Newest Google row to track for this table. It may beat older Pro rows in practical coding/speed but still needs internal hallucination and instruction-following runs.

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 Google release coverage