Skip to content

Commit bf7f766

Browse files
committed
fix(opencode-eval): drop the [1m] suffix from the default model ID
`opencode run --model deepseek/deepseek-v4-pro[1m]` fails fast with: Model not found: deepseek/deepseek-v4-pro[1m]. Did you mean: deepseek-v4-pro, deepseek-v4-flash, deepseek-chat? The `[1m]` suffix is an Anthropic-compat-shim convention for picking Claude's 1M-context variant; OpenCode reads provider/model IDs from Models.dev's deepseek registry, which only exposes the plain IDs. Switch the driver's DEFAULT_MODEL to `deepseek/deepseek-v4-pro` and leave a comment naming the failure mode so a future contributor doesn't paste the `[1m]` suffix back in by reflex. Surfaced by the first dispatch of manual-opencode-judge.yml against ref=opencode-evals — every scenario test failed with the model-not-found error before reaching any skill work. Signed-off-by: Tanzim Saqib <[email protected]>
1 parent 1c13d64 commit bf7f766

1 file changed

Lines changed: 9 additions & 1 deletion

File tree

‎skills-evals/src/skills_evals/opencode_driver.py‎

Lines changed: 9 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -59,7 +59,15 @@
5959
DEFAULT_MAX_TURNS = 20
6060
DEFAULT_PER_TURN_TIMEOUT_S = 600.0
6161

62-
DEFAULT_MODEL = "deepseek/deepseek-v4-pro[1m]"
62+
DEFAULT_MODEL = "deepseek/deepseek-v4-pro"
63+
# Note: the trailing "[1m]" variant suffix that Claude Code uses
64+
# (`deepseek-v4-pro[1m]` — the 1M-context build) is an Anthropic-shim
65+
# convention. OpenCode reads provider/model IDs from Models.dev's
66+
# deepseek registry, which exposes the plain IDs only
67+
# (`deepseek-v4-pro`, `deepseek-v4-flash`, `deepseek-chat`). Adding the
68+
# `[1m]` suffix to the --model flag fails fast with:
69+
# Model not found: deepseek/deepseek-v4-pro[1m]. Did you mean:
70+
# deepseek-v4-pro, deepseek-v4-flash, deepseek-chat?
6371

6472
# Env vars echoed at startup for CI-log clarity. Anything carrying a
6573
# secret reports as set/MISSING with a 4-char suffix, never the value.

0 commit comments

Comments
 (0)