docs: refresh the M3 capability evidence with the 08-25 retest

can_disable stays true — reasoning_effort=none still lands prompt 194,
completion 3, no prose. What the retest added are two limits worth
recording: the verdict is unobservable on the non-streaming path, where
reasoning is billed but neither prose nor usage detail comes back, and
enable_thinking / thinking:{enabled} remain inert on this model.

No behaviour changed, so there is no failing test to show first. The
evidence for a declaration that still holds is the retest itself, not
a unit test the library could write about its own claim.
This commit is contained in:
2026-08-26 00:06:22 -04:00
parent 8c5c23ae72
commit 3e869b9b39
2 changed files with 10 additions and 3 deletions
+3 -2
View File
@@ -72,8 +72,9 @@ DEFAULT_PROFILES: Mapping[str, ProviderProfile] = MappingProxyType(
strip_think_tags=False,
),
# 注入形态出处: 2026-08-02 经自建 new-api 中转实测(findings §2),
# **直连官方端点未验证**。实测 enable_thinking / thinking 两种写法均被
# 静默丢弃(prompt_tokens 恒定不变),reasoning_effort 才是真开关。
# 2026-08-25 复测结论不变(findings 2026-08-25 §5);**直连官方端点未验证**。
# 实测 enable_thinking / thinking 两种写法均被静默丢弃(prompt_tokens
# 恒定等于基线 194),reasoning_effort 才是真开关——本片段的选型据此成立。
# "开"取 medium: qwen 的 enable_thinking:true 与 deepseek 的
# thinking:{enabled} 都不指定预算、由模型自定,medium 是五档里语义最接近
# "厂商正常强度"的一档;取 high 等于替下游做"加钱换质量"的业务判断。
+7 -1
View File
@@ -68,7 +68,13 @@ DEFAULT_CAPABILITIES: Mapping[str, ThinkingCapability] = MappingProxyType(
{
"MiniMax-M3": ThinkingCapability(
can_disable=True,
evidence="2026-08-02 经 new-api 中转实测 N=10: reasoning_effort=none 稳定关闭,零跳变",
evidence=(
"2026-08-02 经 new-api 中转实测 N=10: reasoning_effort=none 稳定关闭,零跳变;"
"2026-08-25 复测依然成立(prompt 194 = 基线、completion 3、无推理正文)。"
"两条限制(findings 2026-08-25-thinking-observability-regression §3.1/§5): "
"① 非流式路径观测不到推理信号——推理已计费,但正文与 usage 明细都不回传;"
"② enable_thinking / thinking:{type:enabled} 对本模型无效,仅 reasoning_effort 是真开关"
),
),
"MiniMax-M2.7": ThinkingCapability(
can_disable=False,