Kwok Ching Yu appeals a sentence imposed by the United States District Court for the Southern District of New York (Kaplan, J.) after Yu pleaded guilty to two counts of conspiracy to distribute heroin and one count of attempted distribution of heroin, in violation of the federal narcotics laws. The district court sentenced Yu to concurrent 20 year terms of imprisonment on each count. On appeal, Yu challenges his sentence primarily on the basis of Apprendi v. New Jersey, 530 U.S. 466, 120 S. Ct. 2348, 147 L. Ed. 2d 435 (2000). Yu also contends that the district court committed clear error in calculating the drug quantity attributable to his crimes, and abused its discretion in declining to award Yu reductions in his offense level based on his purportedly "minor role" in the conspiracy and his purportedly "timely" acceptance of responsibility.

* Yu has been incarcerated since November 1993 at the United States Penitentiary in Lewisburg, Pennsylvania ("Lewisburg"), following his conviction for leading a conspiracy in the 1980s to import heroin into the United States from Hong Kong. The present conviction arises out of a subsequent conspiracy to possess and distribute heroin in which Yu participated, while incarcerated, by giving the names and contact numbers of likely purchasers to a drug dealer named Eric Au Yong.

Au Yong was arrested on narcotics charges in the spring of 1997 and began cooperating with federal law enforcement authorities. He told the government that he had visited Yu in prison in late 1994 or early 1995 in an effort to sell approximately eleven "units" — or 7.7 kilograms1 —of heroin that Au Yong had acquired from China. Au Yong told the government that in the course of several visits, Yu provided him with names and telephone numbers for potential buyers of the heroin, and in at least some cases Yu assured Au Yong that they were reliable people. This information proved fruitful; Yu told Au Yong that he thought Ah Chau would be a good and reliable customer, and in 1996, Au Yong sold Ah Chau a total of 4.5 units — or 3.15 kilograms — in four transactions. At the behest of the government, Au Yong continued to visit Yu at Lewisburg in 1997 and received more information about potential customers. Based on referrals by Yu, Au Yong made more heroin sales: two or three sales of one-half unit each (totaling 700 1050 grams) to one "Chemy"; three ounces (about 85 grams) to one "Jabbar"; two ounces (about 57 grams) to one "Blass"; a few ounces to one "Victor"; and a few ounces to one "Bonnie."

On February 22, 2000, Yu was named in three counts of a four-count indictment. Count One charged Yu and others under 21 U.S.C. § 846 with conspiring to distribute one kilogram and more of heroin between 1995 and 1996, in violation of 21 U.S.C. §§ 812, 841(a)(1) and 841(b)(1)(A). Count Two charged Yu and other defendants with violating Section 846 by conspiring to distribute 100 grams and more of heroin in 1997, in violation of Sections 812, 841(a)(1) and 841(b)(1)(B). Count Three charged Yu and other defendants with violating Section 846 and 18 U.S.C. § 2 by attempting to possess, with intent to distribute, 100 grams and more of heroin, in violation of Sections 812, 841(a)(1) and 841(b)(1)(B).

The case was scheduled for trial on April 11, 2000. On April 5, Yu pleaded guilty to all three counts. Yu did not enter into a plea agreement with the government, and he did not stipulate to the quantity of drugs involved. At the allocution, the district judge read the indictment to Yu and informed him of the penalties he potentially faced, including the possibility of mandatory minimum terms of 20 years on Count One and ten years on Counts Two and Three due to Yu's prior felony conviction. Yu pleaded guilty to each count, but he refused to allocute to the quantities of heroin alleged in the indictment. His attorney explained to the court that Yu intended merely to admit his guilt in conspiring to distribute heroin, with weight to be determined at sentencing because weight "is not an element of the offense."

The Pre-Sentence Report ("PSR") recommended that Yu be sentenced to 240 months in prison. The PSR determined that Yu's offenses should be grouped pursuant to U.S.S.G. § 3D1.2(d), and that Yu should be held accountable for 7.7 kilograms of heroin — the entire amount that Au Yong had been trying to sell with Yu's help. This quantity of heroin yielded a base offense level of 34 under the federal Sentencing Guidelines. The PSR recommended no reduction relating to Yu's role in the conspiracy, but it did recommend a two-level reduction for acceptance of responsibility, resulting in a final offense level of 32. The PSR recommended that Yu be assigned to Criminal History Category III, based on Yu's prior felony conviction and on the fact that he committed the offenses while incarcerated. Based on the offense level of 32 and the Criminal History Category of III, the PSR arrived at a Guideline sentencing range of 151-188 months. The PSR further determined, however, that the drug quantity, because it exceeded one kilogram, triggered the mandatory minimum sentence contained in Section 841(b)(1)(A), which, in light of Yu's prior felony conviction, was 240 months.

At a hearing on December 19, 2000, the district judge heard testimony regarding the drug transactions at the center of the conspiracy and found that Yu was responsible for "substantially in excess" of one kilogram of heroin.

Prior to sentencing, Yu argued in a submission to the court that the mandatory minimum sentence under Section 841(b)(1)(A) could not result from a finding of drug quantity by the district judge, because Apprendi v. New Jersey, 530 U.S. 466, 120 S. Ct. 2348, 147 L. Ed. 2d 435 (2000) — which had been decided in June 2000, two months after Yu entered his guilty plea — barred the application of a mandatory minimum sentence based on a fact not found by a jury beyond a reasonable doubt, where that sentence would exceed the otherwise-applicable sentencing range under the Guidelines. Yu also challenged the PSR's calculation of drug quantity and argued that he should receive additional offense-level reductions for his "minor role" in the conspiracy and because his acceptance of responsibility had been "timely."

On April 9, 2001, the district court sentenced Yu principally to concurrent 240 month terms of imprisonment on each of the three counts. The court found that the weight of heroin exceeded three kilograms, which, combined with other sentencing factors, yielded a Guideline range of 151-188 months. The court further determined, however, that the drug quantity (combined with Yu's prior felony conviction) triggered the mandatory 240-month minimum sentence in Section 841(b)(1)(A), which applies to heroin quantities of one kilogram or more. The district court rejected Yu's Apprendi argument on the ground that Apprendi had left intact the Supreme Court's prior holding in McMillan v. Pennsylvania, 477 U.S. 79, 106 S. Ct. 2411, 91 L. Ed. 2d 67 (1986), that judicial findings of fact made under a preponderance standard may trigger mandatory minimum sentences. The court also rejected Yu's other arguments, finding that Yu was not entitled to a minor-role adjustment and that an offense-level reduction for "timely" acceptance of responsibility was unwarranted given that Yu had "waited until the last second to plead."

On appeal, Yu presses his argument under Apprendi, challenges the district judge's finding as to drug quantity, argues that the district court abused its discretion in declining to view Yu's role in the conspiracy as "minor" (thereby denying him a two-point offense-level reduction), and argues that the district court abused its discretion in declining to view Yu's acceptance of responsibility as "timely" so as to warrant an additional one-level reduction beyond the two-level reduction Yu had already received.

Yu's Apprendi argument is the only one with any merit. We address it first, then turn briefly to his other arguments.

Yu refused to allocute to drug quantity, and that determination ultimately was made by the district judge at sentencing under a preponderance standard. Because drug quantity was not found by a jury beyond a reasonable doubt, Yu argues that he had to be sentenced under Section 841(b)(1)(C), the provision applicable to narcotics offenses absent the appropriate finding as to quantity. Because that section contains no mandatory minimum sentence, Yu argues that he should have been sentenced within the range specified by the Guidelines (which the district court determined to be 151-188 months, though Yu also disputes this, see infra). Instead, the district court invoked Section 841(b)(1)(A) and sentenced Yu to the 240-month mandatory minimum sentence prescribed in that subsection. Yu argues that because the court's application of subsection (A) (and, thus, his increased sentence) was triggered by a fact — drug quantity — that the court found under a preponderance standard, this was error under the rule established in Apprendi v. New Jersey, 530 U.S. 466, 120 S. Ct. 2348, 147 L. Ed. 2d 435 (2000), that any fact, other than a prior conviction, that increases the penalty for a crime beyond the prescribed statutory maximum must be submitted to a jury and proved beyond a reasonable doubt.

Yu's argument was made prior to, and anticipates, this Court's recent decision in United States v. Thomas, 274 F.3d 655 (2d Cir.2001) (in banc). There, we held that Apprendi applies to findings of drug quantity under the federal narcotics statute, 21 U.S.C. § 841 (2000). 274 F.3d at 660. That statute provides escalating penalties for narcotics offenses depending chiefly on the quantities of narcotics involved. Applying Apprendi, we held in Thomas that where drug quantity is not found by a jury beyond a reasonable doubt, but rather is determined by the district court at sentencing under a preponderance standard, the defendant must be sentenced under 21 U.S.C. § 841(b)(1)(C), the subsection applicable to narcotics offenses without regard to quantity. Id. at 663, 667. Thus, Yu's argument that he should have been sentenced under Section 841(b)(1)(C) because drug quantity was determined by the district judge under a preponderance standard is facially consistent with our holding in Thomas.

Because Yu's sentence did not exceed the 20 year maximum in Section 841(b)(1)(C), his argument also depends on the proposition that Apprendi is violated when a judge's finding triggers a mandatory minimum sentence that is less than the maximum penalty provided for by the penal statute but that exceeds the otherwise-applicable sentencing range under the federal Guidelines. Shortly after we decided Thomas, we adopted this view in United States v. Guevara, 277 F.3d 111 (2d Cir. 2001).

In this case, however, the reason drug quantity was never submitted to the jury is that Yu pleaded guilty, waiving his right to jury determination of any issue. At his plea hearing, Yu explicitly stated his willingness to have the district judge determine drug quantity at sentencing because, as Yu's attorney acknowledged, drug quantity was not an element of a Section 841 offense. All of this happened, of course, before Apprendi and Thomas, which together require that drug quantity be treated as an element — i.e., be found by a jury beyond a reasonable doubt — in order for a defendant to be sentenced for a Section 841(a) offense under any subsection other than § 841(b)(1)(C). In light of his counsel's comment, it seems clear that Yu's decision to submit the drug-quantity question to the district judge under the preponderance standard was influenced by his understanding — at that time — that he would not have been entitled to a jury determination of that issue in any event.

We agree with Yu that there is a defect in the proceedings: in the light (or hindsight) of Apprendi, Thomas and Guevara, it was error for the district court to permit Yu to plead guilty to quantity-specific charges while refusing to allocute to quantity. But because Yu pleaded guilty, the nature of the defect is somewhat different and the question of relief is different altogether. Yu pleaded guilty to an indictment that charged specific quantities of narcotics and that made specific reference to Sections 841(b)(1)(A) and 841(b)(1)(B). We can assume that the government was prepared to offer evidence as to quantity at trial. Without challenging or seeking to withdraw his plea, however, Yu asks us to amend it by holding that he really pleaded guilty to, and was thus convicted of, a Section 841(b)(1)(C) offense rather than the more severe offenses with which he was actually charged.

We need not consider whether Yu would be entitled to withdraw his plea; that is a matter largely left to the discretion of the district court. See, e.g., United States v. Hernandez, 242 F.3d 110, 112 (2d Cir.2001) (noting that a district court's denial of a motion to withdraw a guilty plea is reviewed for abuse of discretion); United States v. Avellino, 136 F.3d 249, 261 (2d Cir.1998) ("As a general matter, a defendant has no absolute right to withdraw [a] plea of guilty."). Ordinarily, the failure to seek that relief would be a waiver, but the circumstances here are unusual (to say the least) and the government has not invoked the doctrine, so it may be that any claim of waiver has itself been waived. See Tischmann v. ITT/Sheraton Corp., 145 F.3d 561, 568 n. 4 (2d Cir.1998) (noting that a failure to timely assert an argument on appeal constitutes waiver).

On remand, the district court should be afforded the opportunity and the flexibility to consider and order appropriate relief. It may be that Yu will seek to withdraw his plea; it may be that the government in that event would prefer to acquiesce in a re-sentencing within the Guideline range applicable to the offense sufficiently described in Yu's original allocution; it may be that Yu will seek a jury trial on the issue of drug quantity; it may be that the government would in that event argue that Yu's waiver of a jury trial in the plea allocution was a voluntary and knowing waiver of a jury trial on the issue of drug quantity2; and it may be that Yu will seek or be consigned to a bench trial at which a finding on quantity would be made beyond a reasonable doubt. These non-exclusive alternatives (depending on the district court's findings, rulings, and discretion) are avenues of relief for the Apprendi problem that has arisen in Yu's case. But Yu is not entitled to plead guilty to an indictment charging a violation of Sections 841(b)(1)(A) and (B) and then be sentenced under Section 841(b)(1)(C), as he would have been if a jury had decided the drug-quantity issue in his favor.

Accordingly, we vacate Yu's sentence and remand this case to the district court for further proceedings consistent with this opinion. Any sentence under subsections (A) or (B) must be based either on an allocution that settles the issue of drug quantity or on a finding as to that issue by a fact-finder applying a reasonable-doubt standard. Whether that means allowing Yu to withdraw or modify his plea or to proceed directly to a trial on the issue of drug quantity alone, or whether it means some other course of action, are questions we confide to the judgment and discretion of the district court.

On the assumption that his Apprendi argument leads to a re-sentencing, Yu argues next that the district court arrived at the wrong Guideline range (151-188 months) due to erroneous findings at sentencing. (We consider these arguments because we have held supra that Apprendi is implicated in this case, even though the proceedings on remand may not result in a re-sentencing.) Specifically, Yu contends that the court erred in calculating the drug quantity attributable to Yu's crimes, that it abused its discretion in denying him a two-point reduction in his offense level for playing a minor role in the conspiracy, and that it abused its discretion in denying him a one-point reduction for accepting responsibility in a timely fashion.

A district court's findings of fact at sentencing are reviewed for clear error. United States v. Hazut, 140 F.3d 187, 190 (2d Cir.1998). A finding of fact is clearly erroneous only if the appellate court is left "with the definite and firm conviction that a mistake has been committed." Anderson v. Bessemer City, 470 U.S. 564, 573, 105 S. Ct. 1504, 84 L. Ed. 2d 518 (1985) (quoting United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S. Ct. 525, 92 L. Ed. 746 (1948) (internal quotation marks omitted)).

Yu contends that the district judge erred in holding him responsible for the 3.15 kilograms of heroin that Au Yong sold to Ah Chau, and that the amounts sold to the other buyers do not add up to the threshold required for a base offense level of 34.

We see no basis — under any standard, much less the clear error standard — for overturning the district judge's finding that Yu was responsible for the amount sold to Ah Chau. Yu's theory is that he should not be held liable for the transaction because Au Yong and Ah Chau were already acquainted with one another, and Yu "did not set up or facilitate meetings or sales between the two men; he was not even aware of at least two of the deals between the men; nor did he profit from the transactions."

Yu ignores the evidence from which the district court concluded that he had essentially brokered the deal between Au Yong and Ah Chau. Au Yong testified that Ah Chau had expressed an interest in purchasing heroin from him, but that Ah Chau had just been released from jail and Au Yong consequently "didn't have much confidence in him." He testified further that he sought Yu's advice because he knew Yu was a friend and partner of Ah Chau. During the prison visit, Yu promised Au Yong that he would contact Ah Chau, and at their next meeting he told Au Yong that he had spoken with Ah Chau and felt "that there is no problem." Relying on this assurance, Au Yong sold a total of 3.15 kilograms of heroin to Ah Chau on four different occasions. This testimony furnished sufficient basis for a finding that Yu had brokered the deal between his two acquaintances and consequently was liable as a co-conspirator — or (alternatively, as the government points out) as an aider and abettor. See, e.g., U.S. SENTENCING GUIDELINES MANUAL § 1B1.3, ill. (a)(1) (2001). It is not decisive that Yu did not introduce the participants in the drug transaction, or that he may have been unaware of the third and fourth transactions; he was plainly aware that Au Yong was trying to sell units and half-units of heroin, since he was aware that Au Yong twice sold half-units to Ah Chau. Thus, as the district judge found, it was certainly "overwhelmingly foreseeable" to Yu that as much as 3.15 kilograms of heroin might change hands between the two.

Yu argues next that the district court abused its discretion in finding that Yu failed to carry his burden of showing, by a preponderance of the evidence, that he had a "minor role" in the conspiracy and thus merited a two-point reduction in his offense level.

In evaluating a defendant's role, we look to factors such as "the nature of the defendant's relationship to other participants, the importance of the defendant's actions to the success of the venture, and the defendant's awareness of the nature and scope of the criminal enterprise." United States v. Garcia, 920 F.2d 153, 155 (2d Cir.1990) (upholding district court's refusal to grant minor-role reduction to a drug courier). As noted in Garcia, this determination is fact-sensitive. Id. Moreover, this Circuit has held that a minor-role adjustment is not available merely on a showing that the defendant "played a lesser role than his co-conspirators; to be eligible for a reduction, the defendant's conduct must be `minor' ... as compared to the average participant in such a crime." United States v. Rahman, 189 F.3d 88, 159 (2d Cir.1999).

Yu supplied Au Yong with multiple names and phone numbers and spoke to potential customers on Au Yong's behalf. Yu's vouching for Ah Chau removed an obstacle to the scheme, at least in Au Yong's eyes. Stated in Garcia terms, Yu's importance relative to the other conspirators was clearly quite high: he was a trusted authority, as it was on his word that the sales to Ah Chau were consummated; he was also aware of the nature and scope of the enterprise, as evidenced (for example) by taped conversations in which Yu suggested to Au Yong that successful sales to co-defendant Lawson could lead to a "deluge" of business.

We therefore agree with the district court that Yu's role in the conspiracy was not "minor" so as to warrant a reduction in his offense level.

The district court granted a two-point reduction in offense level pursuant to U.S.S.G. § 3E1.1(a) based on its finding that Yu had accepted responsibility for his actions. Yu argues, however, that the district court abused its discretion in declining to find that the acceptance of responsibility was "timely" so as to warrant an additional one-point reduction. We disagree. Yu entered his plea barely one week before his trial was to begin; his belated plea was not sufficiently timely to "permit[ ] the government to avoid preparing for trial and . . . the court to allocate its resources efficiently." U.S.S.G. § 3E1.1(b)(2). Moreover, the record shows that the district court detected little of the sincere remorse that is required even for the basic two-level reduction, and granted that reduction only grudgingly, stating "this is as close as I have ever come to denying it." We see no reason to hold that the district judge had to grant Yu an even greater reduction for eleventh-hour contrition the sincerity of which was doubtful to begin with.

Finally, the district court committed a sentencing error that the government brings to our attention. Counts Two and Three of the indictment alleged conspiracy to distribute 100 grams and more of heroin and attempted distribution of 100 grams and more of heroin, respectively. On both counts the court sentenced Yu to the mandatory minimum term of 240 months under Section 841(b)(1)(A). As the government acknowledges, however, Counts Two and Three charge a threshold quantity of heroin—100 grams and more — that constitutes a violation only of Section 841(b)(1)(B), not subsection (A) (for which the threshold is one kilogram and more). Accordingly, the district court should not have sentenced Yu on Counts Two and Three to the 240-month mandatory minimum term under subsection (A). Rather, Yu should have been sentenced on those counts under subsection (B), for which the mandatory minimum term is only 120 months. On the government's view, Yu should be sentenced to the Guideline range of 151-88 months; we do not decide this question.

Without knowing what may come of further proceedings in the district court, we are not in a position to mandate how Yu should be sentenced on Counts Two and Three. At this stage, all we can say is that it was error to apply a subsection (A) penalty to the subsection (B) offenses charged in Counts Two and Three of the indictment.

* * *

For the reasons stated, the sentence is hereby vacated and the case is remanded for further proceedings, including vacatur of the judgment as may be necessary, consistent with this opinion.