agentic-workspace live run — 2026-08-26

Model: openai/gpt-oss-20b

Driver: in-process · Context length: 16384

Scenario	Result	Checks	Tool calls	Rounds	Wall (s)	Tokens (prompt/predicted)
S1	✅	6/6	4	5	2.3	15317/262
S3	✅	11/11	8	10	4.5	44565/684
S5	✅	3/3	4	5	2.4	16454/368
S9	❌	11/11	14	15	9.8	116370/1704

Failures

S9 — full project from an empty directory (context overflow: prompt 16931 > 16384)

⚠️ = the run reached the context ceiling. Where the row still passes, the scored work had already completed and been written to disk before the model ran out of context; a longer task at the same context length would be cut off mid-work. Pair the plugin with a context compressor, or raise the context length.

Pass = every check ok. Checks assert filesystem, git and .agentic state, never model prose. Results are comparable only within one model.

agentic-workspace live run — 2026-08-26

Model: openai/gpt-oss-20b

Driver: in-process · Context length: 16384

Scenario	Result	Checks	Tool calls	Rounds	Wall (s)	Tokens (prompt/predicted)
S1	✅	6/6	4	5	2.3	15317/262
S3	✅	11/11	8	10	4.5	44565/684
S5	✅	3/3	4	5	2.4	16454/368
S9	❌	11/11	14	15	9.8	116370/1704

Failures

S9 — full project from an empty directory (context overflow: prompt 16931 > 16384)

Pass = every check ok. Checks assert filesystem, git and .agentic state, never model prose. Results are comparable only within one model.