FUSTON v. QUICK (2026)

Ronnie Fuston was tried and convicted for the murder of Michael Rhodes. Fuston broke into Mr. Rhodes's home late at night and shot him three times while he was seated on a couch next to his three-year-old daughter. Mr. Rhodes, it turned out, was an innocent victim caught in the middle of a gang dispute.

An Oklahoma jury sentenced Fuston to death. On direct appeal to the Oklahoma Court of Criminal Appeals (OCCA), he challenged the trial court's failure to give a lesser-included instruction on second-degree murder and his death sentence. After the OCCA affirmed his conviction and sentence, he petitioned for federal habeas relief, arguing that his conviction and sentence violated his constitutional rights because (1) due process entitled him to an instruction on the lesser-included offense of second-degree murder under Beck v. Alabama, 447 U.S. 625 (1980), and (2) the trial court could not impose the death penalty without an evidentiary hearing to determine whether he was intellectually disabled under Atkins v. Virginia, 536 U.S. 304 (2002). The district court denied Fuston's habeas petition, and we granted a certificate of appealability on both issues. See 28 U.S.C. § 2253(c)(1)(A).

Exercising our jurisdiction under 28 U.S.C. § 1291, we AFFIRM the denial of Fuston's habeas petition. Fuston seeks relief under 28 U.S.C. § 2254(d)(1), which permits a court to grant the writ only if the state court's determination was “contrary to, or involved an unreasonable application of” law clearly established by the Supreme Court. He argues the OCCA unreasonably applied both Beck and Atkins. We disagree.

First, Beck requires that the trial court, in capital cases, instruct the jury on a lesser-included offense only when the evidence supports the offense. The OCCA's determination that the evidence would not have permitted a rational juror to find Fuston guilty of second-degree murder and acquit him of first-degree murder was a reasonable evaluation of the facts presented at trial. The OCCA therefore did not unreasonably apply Beck.

Second, under Oklahoma law a defendant is not considered intellectually disabled if he presents an IQ score above 75. Okla. Stat. tit. 21, § 701.10b(c) (2006). Although Fuston presented scores both above and below 75, applying Oklahoma law, the OCCA determined that Fuston's score of 81 precluded a finding that he was intellectually disabled. It then affirmed the trial court's denial of an evidentiary hearing on the intellectual-disability issue. We find that Atkins and its progeny do not clearly establish that a defendant who presents multiple IQ scores—one above 75; the others, below—must receive an Atkins hearing. The OCCA therefore did not unreasonably apply clearly established law when it affirmed Fuston's death sentence.

I. Background 1

In October 2012, Ronnie Fuston joined with members of a gang and entered Michael Rhodes's home. Once inside, Fuston shot and killed Rhodes while he lay on a couch next to his three-year-old daughter.

As it happens, Rhodes was an innocent victim caught in the crossfire of a gang dispute. The dispute arose from animosity between two female gang members—Atiana Jordan and Taneecia Pennon—and Rhodes's niece, Brittany Dillard. Dillard resided at the Rhodes home before the shooting. Jordan and Pennon made threatening calls to the Rhodes household, threw rocks into the home's windows, and slashed the tires of a vehicle parked in the driveway.

The gang members called Fuston to enlist his help in resolving the dispute. He agreed and drove to Oklahoma City to meet the gang members. He and several other gang members then drove to the Rhodes home, where they wrongly believed Dillard was staying. (In fact, she had relocated to her boyfriend's home sometime before the shooting.) As the group kicked the door down, Fuston began shooting “almost simultaneously.” Tr. 953. He shot Rhodes three times.

After shooting Rhodes, Fuston and the gang members left the home and returned to the car. But Jordan hesitated to get in the car to leave because she believed the job was not finished; she said they were supposed to kill everyone in the house. And once Fuston was in the car, his friend, Brian Butler, asked him why he shot twice. Fuston corrected him: because Rhodes started moving, he shot at Rhodes four times, not two.

The State charged Fuston with first-degree murder. At trial, Fuston requested an instruction on second-degree murder. He argued that the evidence supported a finding that he shot into the home “with a depraved mind, with a reckless disregard for life, but without the intent to kill.” Tr. 1643. To support that instruction, he pointed to evidence that (1) the “kicking of the door and the firing on the home came fairly simultaneously” and but for “the fact that the homeowner was getting up[,] the shots would have hit the fireplace and not him” and (2) the “gang girls” intended to vandalize the home, not “to kill anyone.” Tr. 1643. The court denied the instruction.

Fuston also requested an evidentiary hearing to determine his intellectual disability under Atkins. Fuston presented IQ scores of 59, 69, 75, and 81. Focusing on the 81 score, the court denied the request for an Atkins hearing because Oklahoma law mandates that one whose IQ score is “seventy-six (76) or above” cannot “be considered intellectually disabled.” Okla. Stat. tit. 21, § 701.10b(c) (2006).²

The jury convicted Fuston of first-degree murder. It also found two aggravating circumstances justifying the death penalty: (1) Fuston “knowingly created a great risk of death to more than one person”, and (2) he would probably “commit criminal acts of violence that would constitute a continuing threat to society.” See Okla. Stat. tit. 21, § 701.12 (2006).³ Fuston was sentenced to death.

Fuston appealed. Relying on Beck, he argued that the court violated his due-process rights by refusing to give an instruction on the lesser-included offense of second-degree murder. The OCCA affirmed. It focused on the intent element of second-degree murder—which requires an absence of “premeditated design,” Okla. Stat. Ann. tit. 21 § 701.8 (1976)—and concluded that no evidence would support a finding that Fuston lacked premeditation. Fuston also raised a challenge under Atkins and its progeny, arguing that his death sentence violated the Eighth Amendment because the trial court denied him an Atkins hearing. The OCCA disagreed. It concluded that Oklahoma law requires Fuston's IQ score of 81 be treated as conclusive evidence that Fuston was not intellectually disabled. The OCCA distinguished Atkins and Hall, concluding those cases did not address a fact pattern where the defendant presents an IQ score well above the high end of the standard error of measurement (SEM) along with scores below the low end. The OCCA assumed that Fuston's IQ tests had a SEM of five points. See Fuston, 470 P.3d at 316 (“The SEM for IQ tests is generally considered to be plus or minus five points.” (citing Hall, 572 U.S. at 713)).⁴

While Fuston's direct appeal was pending, he filed for post-conviction relief with the OCCA. And after the OCCA affirmed on direct appeal, it denied his request for post-conviction relief. Fuston petitioned for certiorari, and the Supreme Court denied the petition.

Fuston then filed for habeas relief in federal court. Applying the required deference to state-court determinations, the district court denied Fuston's habeas petition. The court concluded that the state court's failure to give an instruction on second-degree murder was not an unreasonable application of Beck because the evidence at trial did not support a finding that Fuston lacked premeditated intent. The court also concluded that denying an Atkins hearing was not contrary to clearly established law because Atkins and its progeny did not address scenarios with multiple IQ scores. The court therefore denied the habeas petition and denied a COA.

Fuston moved this court for a COA, and we granted on two issues: (1) whether the state court's failure to instruct on second-degree murder violated his due-process rights, and (2) whether imposition of the death penalty without an Atkins hearing violated Fuston's Eighth Amendment rights because he is intellectually disabled.

II. Discussion

We affirm the district court on both grounds because the OCCA did not unreasonably apply either Beck or Atkins.

A. Standard of Review under AEDPA

Federal law authorizes federal district courts to issue writs of habeas corpus. 28 U.S.C. § 2241. And the Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA), 110 Stat. 1214 (codified throughout at 28 U.S.C. §§ 2244–2255), circumscribes the writ's availability to challenge state custody. See 28 U.S.C. § 2254. Most relevant here is § 2254(d)(1), which allows habeas relief only if the state court's “adjudication of the claim” was “contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States.”  ⁵

The threshold question is therefore whether the legal proposition claimed by the petitioner is clearly established by Supreme Court precedent. House v. Hatch, 527 F.3d 1010, 1014–15 (10th Cir. 2008). If the applicant fails to make that showing, the reviewing court need not assess whether the state court's adjudication was contrary to, or unreasonably applied, the legal proposition. Littlejohn v. Trammell, 704 F.3d 817, 825 (10th Cir. 2013). The law must be clearly established “at the time the state conviction became final.” House, 527 F.3d at 1015 (quoting Williams v. Taylor, 529 U.S. 362, 380 (2000)).

Only holdings of the Supreme Court clearly establish legal propositions for AEDPA purposes. See 28 U.S.C. § 2254(d)(1). The Supreme Court's holdings are those “legal rule[s] or principle[s]” that the Court “relies on ․ to decide a case.” Andrew v. White, 604 U.S. 86, 92 (2025). And so long as the Supreme Court relies on them, even “[g]eneral legal principles can constitute clearly established law for purposes of AEDPA.” Id. Still, a legal proposition is not clearly established if it raises a question explicitly left open by the Supreme Court. See Carey v. Musladin, 549 U.S. 70, 76–77 (2006) (finding proposition not clearly established because the “Court ha[d] never addressed a claim” raising the issue before the Court); Paxton v. Ward, 199 F.3d 1197, 1206 (10th Cir. 1999) (finding law not clearly established because “[t]he Supreme Court has expressly left open” the issue).

If the applicant's asserted legal proposition is clearly established, then the reviewing court must assess whether the state court's adjudication was “contrary to” or was an “unreasonable application” of the clearly established law. 28 U.S.C. § 2254(d)(1). These clauses have “independent meaning.” Williams v. Taylor, 529 U.S. 362, 405 (2000). A state-court decision is contrary to federal law if it either (1) “ ‘arrives at a conclusion opposite to that reached by’ the Supreme Court” or (2) “ ‘decides a case differently’ than the Court on a ‘set of materially indistinguishable facts.’ ” Wood v. Carpenter, 907 F.3d 1279, 1289 (10th Cir. 2018) (quoting Williams, 529 U.S. at 412–13). A state-court decision unreasonably applies federal law if it “ ‘identifies the correct governing legal principle’ ․ but applies [that] principle[ ] in an objectively unreasonable manner.” Id. (quoting Wiggins v. Smith, 539 U.S. 510, 520 (2003)). An application is objectively unreasonable only if “all fairminded jurists would agree the state court decision was incorrect.” Id. (quoting Frost v. Pryor, 749 F.3d 1212, 1225 (10th Cir. 2014)). A state court's factual findings are presumptively correct unless rebutted by clear and convincing evidence. 28 U.S.C. § 2254(e)(1).

B. Beck Error

Fuston charges the OCCA with unreasonably applying Beck v. Alabama, 447 U.S. 625 (1980). Beck requires that, in capital cases, the judge instruct the jury on a lesser-included offense “when the evidence warrants such an instruction.” Hopper v. Evans, 456 U.S. 605, 611 (1982) (discussing Beck).⁶ Beck challenges have some unique features when subject to AEDPA deference. For instance, we will not defer to a state court's Beck determination if it applies the wrong standard. Accordingly, we do not defer when the state court determines that the evidence is consistent with the greater offense rather than whether the evidence would support the lesser offense. See Taylor v. Workman, 554 F.3d 879, 888 (10th Cir. 2009); Darks v. Mullin, 327 F.3d 1001, 1009–10 (10th Cir. 2003) (explaining that denying a manslaughter instruction because sufficient evidence “support[ed] a finding of premeditation in the trial record is squarely contrary to the holding in Beck”). In this scenario, the circuit “engage[s] in de novo review” to determine whether the “evidence was sufficient to warrant” an instruction on the lesser-included offense. Phillips, 604 F.3d at 1213.

But if the state court applies the correct rule and asks whether the evidence supports the lesser-included offense, the reviewing court asks whether the state court unreasonably applied Beck. See Richie v. Workman, 599 F.3d 1131, 1140–41 (10th Cir. 2010) (holding that a state court's conclusion that evidence did not support second-degree murder “was an unreasonable application of” Beck). And under the unreasonable-application rubric, the state court's application of Beck—specifically, its conclusion that the evidence did not support the lesser-included instruction—is unreasonable only if “all fair-minded jurists would agree that the state court decision was incorrect.” Frost, 749 F.3d at 1225.

Fuston does not clearly say whether we owe deference to the OCCA's determination. In his briefing, he represented that the OCCA's determination does “not warrant this [c]ourt's deference” because the OCCA performed the wrong Beck analysis. See, e.g., Appellant Br. on Ground IV at 49. But at oral argument, he conceded that we do owe deference to the OCCA's determination and that the OCCA stated the correct standard for the Beck analysis. Oral Argument, 2:35–2:47 (owes deference); 5:15–5:25 (correct standard). Nonetheless, he opined that the OCCA's decision here was similar to other Beck cases in which we did not defer to the OCCA because it applied the wrong standard. See, e.g., Hogan v. Gibson, 197 F.3d 1297, 1306 (10th Cir. 1999) (giving no deference to the OCCA because it “made no findings as to whether” the evidence was sufficient to warrant a lesser-included instruction).

Regardless of Fuston's apparent concession that we owe deference to the OCCA's determination, we independently find that AEDPA requires deference. The OCCA applied the correct standard under Beck because it properly considered—and determined—that the evidence would not permit a rational juror to convict on second-degree murder while acquitting on first-degree murder. See Hogan, 197 F.3d at 1305 (setting out correct Beck standard). The OCCA explained that Fuston “did not present any evidence, nor did the State's case provide any, that showed he engaged in imminently dangerous conduct in extreme disregard for human life without the intent of taking the decedent's life.” Fuston, 470 P.3d at 325 ; see also id. at 325–26 (rejecting Beck challenge because Fuston “has not shown that the evidence presented at trial would permit a rational jury to find him guilty of second degree depraved mind murder”). That is the correct inquiry mandated by Beck. See Richie, 599 F.3d at 1136 (explaining that Beck requires a lesser-included instruction if the evidence supports the lesser-included offense).

To be sure, the OCCA found that the “evidence clearly supports a finding that when Appellant fired at the decedent, he did so with the intent to kill.” Fuston, 470 P.3d at 325. And relying on our prior cases setting out the correct Beck standard, Fuston argues that the OCCA could not “focus on evidence of first-degree murder,” Appellant Br. on Ground IV at 10, and should have apparently only looked at the evidence of second-degree murder, Appellant Br. on Ground IV at 24. But Fuston overstates our prior cases. In those cases, we concluded that the OCCA applied the wrong standard because it only looked at whether the evidence was sufficient for first-degree murder. The OCCA's invocation of the evidence of intent was not simply a sufficiency-of-the-evidence analysis for first-degree murder.

The best reading of the OCCA's order is that it concluded (1) Fuston presented no evidence supporting second-degree murder, and (2) the evidence negating an element of second-degree murder—i.e., intent—was so overwhelming that no rational juror could find second-degree murder. See Eizember v. Trammell, 803 F.3d 1129, 1145–46 (10th Cir. 2015) (“[I]ntent to kill does preclude a conviction for ‘depraved mind’ murder under [Oklahoma] law—compelling instead a first-degree murder conviction ․”); Taylor, 554 F.3d at 891 (concluding that “evidence regarding intent” was not “so one-sided that a reasonable jury had no alternative but to conclude” the defendant acted with “premeditated design”). But even assuming the second rationale rested on the wrong standard, the first rationale employed the correct one. That is enough to conclude the OCCA applied the correct standard. See Gilson v. Sirmons, 520 F.3d 1196, 1236 (10th Cir. 2008) (rejecting Beck challenge because one of the OCCA's rationales, “standing alone, represents a reasonable application of Beck,” even if the other rationales were “suspect”).⁷ And because the OCCA's decision is not contrary to Beck, we employ the unreasonable-application framework, which entails deference to the OCCA's determination.

Because Fuston challenges the reasonableness of the OCCA's application of Beck to his case, we owe the OCCA's determination substantial deference. To grant Fuston relief, we must conclude that all fair-minded jurists would find the OCCA's application of Beck unreasonable. Fuston must therefore face AEDPA's “formidable barrier to federal habeas relief.” Burt v. Titlow, 571 U.S. 12, 16 (2013).

To make out a Beck claim on second-degree murder, Fuston must show that (1) second-degree murder is a lesser included offense of first-degree murder and (2) the “evidence presented at trial would permit a rational jury to find him guilty” of second-degree murder. Phillips v. Workman, 604 F.3d 1202, 1210 (10th Cir. 2010).

When assessing the sufficiency of the evidence, “courts are not directed to evaluate the evidence to determine whether it would support a first degree murder conviction, or even whether a conviction for first degree murder or a lesser-included offense is better supported.” Phillips, 604 F.3d at 1212 (quoting Taylor, 554 F.3d at 887). The court must limit its inquiry to whether the evidence would support a conviction on the lesser-included offense. The state court must give the instruction “where the evidence would support conviction for first degree murder but would also support conviction on the lesser-included offense.” Taylor, 554 F.3d at 887.

Fuston satisfies the first requirement. He requested an instruction on second-degree depraved-mind murder, which is a lesser-included offense of first-degree murder under Oklahoma law. See Bench v. State, 431 P.3d 929, 954 (Okla. Crim. App. 2018).

The crux of Fuston's challenge therefore falls on Beck’s second requirement. At the time of Rhodes's murder, first-degree murder under Oklahoma law required “deliberate intent to end human life, which can be instantly formed and inferred from the fact of the killing.” Simpson v. State, 230 P.3d 888, 896–97 (Okla. Crim. App. 2010) (quoting Jones v. State, 134 P.3d 150, 154 (Okla. Crim. App. 2006)). And under Oklahoma law, “[a] design to effect death is inferred from the fact of killing, unless the circumstances raise a reasonable doubt whether such design existed.” Taylor, 554 F.3d at 890 (citing Okla. Stat. tit. 21 § 702 (2006)). Second-degree depraved-mind murder, on the other hand, is killing “by an act imminently dangerous to another person and evincing a depraved mind, regardless of human life, although without any premeditated design to effect the death of any particular individual.” Fuston, 470 P.3d at 325 (quoting Okla. Stat. tit. 21 § 701.8(1) (2006)). The relevant difference between those crimes is intent. See id. (“Second degree depraved mind murder is applicable where there is no premeditated intent to kill any particular person.”). Thus, the relevant question under Beck is whether the evidence at trial would permit a rational juror to find that Fuston killed Rhodes without premeditated intent. Applying Beck, the OCCA answered no.

That application of Beck was reasonable. We cannot conclude that all fairminded jurists would think the evidence sufficient to find Fuston lacked intent to kill. Frost v. Pryor, 749 F.3d 1212, 1226 (10th Cir. 2014). We start first with the OCCA's determination that the evidence showed Fuston shot at Rhodes “not twice but four (4) times because he moved.” Fuston, 470 P.3d at 325. At trial, Butler testified that he asked Fuston why he shot into the doorway twice. Fuston answered that he “didn't shoot twice, he shot four times” and that “the guy was getting up or reaching for something, so that's why he shot.” Tr. 1052. Fuston thus shot Rhodes after seeing him get up or reach for something. And critically, Fuston does not dispute that fact;⁸ he disputes only the inference we can draw from it.

The inference he advocates for is speculative. He says a jury could have interpreted the “testimony that Mr. Fuston said he had shot multiple times because he saw someone moving” as evidence “of a reflexive, startled lack of premeditation.” Appellant Br. on Ground IV at 36. But that interpretation of the evidence builds inference upon inference: he invites speculation that (1) Rhodes's movement startled him, and (2) once startled, Fuston reflexively shot at Rhodes without intending to kill him. Fuston points to no evidence supporting the first inference. He does not, for example, point to his own testimony or witness testimony that he was startled. See United States v. Rainford, 161 F.4th 648, 670 (10th Cir. 2025) (requiring involuntary-manslaughter instruction because defendant pointed to testimony about his subjective beliefs, which supported imperfect defense of another); United States v. Pearson, 203 F.3d 1243, 1271 (10th Cir. 2000) (noting that defendant showed lack of intent through witness testimony “that he accidentally fired the shot” that killed the victim). And even if he did show evidence for the first inference, the first inference does not imply the second: Fuston could have been startled by movement and then intended to kill Mr. Rhodes. He offered no evidence to support either of his preferred inferences. And we require more than a chain of speculative inferences to find Beck error. See Darks v. Mullin, 327 F.3d 1001, 1011 (10th Cir. 2003) (finding no Beck error because “any inference of provocation is mere speculation”); Grant v. Trammell, 727 F.3d 1006, 1014 (10th Cir. 2013) (rejecting Beck challenge because [the defendant] did not “point to any facts suggesting (let alone clearly compelling the inference)” he was provoked).

Fuston tries to analogize to Taylor v. Workman, 554 F.3d 879 (10th Cir. 2009), where we held that a defendant was entitled to an instruction on second-degree murder. There, the defendant saw “movement out of the corner of his eye” and started shooting toward the victim. But we—unlike the court in Taylor ⁹ —must defer to the OCCA's finding. Still more troubling for Fuston is that the evidence in Taylor suggested more than that the defendant saw movement and started shooting; the defendant himself testified that he “didn't realize what had really happened at that point and ․ flipped out and started to run out the door.” Taylor, 554 F.3d at 891. Then, when Taylor saw somebody “out of the corner of [his] eye,” he “did not aim as he shot, but the gun was just flailing around.” Id. And when asked whether he shot the victim to kill all witnesses of the crime, Taylor testified, “I don't think I was thinking at all.” Id. That evidence of Taylor's unthinking, aimless shooting, we said, supported a second-degree murder instruction. We do not have that evidence here; we have Fuston saying he shot because Rhodes moved. Deferring as we must to the OCCA's determination, we cannot say that all fair-minded jurists would conclude that shooting someone because he started standing up or reaching for something supported a finding of a lack of intent.

Fuston next says the evidence would have supported a finding that when he fired through the doorway, he fired blindly and could not have known Rhodes was behind the door; Rhodes, in other words, “just so happen[ed]” to be in the line of fire. Fuston, 470 P.3d at 325. Fuston points to the following evidence:

• Testimony from the crime scene investigator that bullet holes were scattered around the crime scene. Tr. 882–84, 895. One in the door near the deadbolt; one in a chair; one near the fireplace. Fuston says this permits an inference that he was not aiming when he fired.

• Testimony from a crime scene investigator that there was no evidence that “anybody that had been involved had ever stepped over the threshold or into the residence.” Tr. 903.

• Testimony from a crime scene reconstructor that “based off of the position” Rhodes was in when shot, “the shooting came almost simultaneously or immediately after the door was kicked.” Tr. 953.

• Testimony from the medical examiner that she “found three separate wounds” in Rhodes's “shoulder, leg and foot.” Tr. 1373. Fuston says that permits an inference that two bullets missed entirely and two were not close to vital organs. That in turn suggests that Fuston was firing blindly.

But even if Fuston did not initially see Rhodes, the uncontroverted evidence showed that Fuston shot at Rhodes after seeing him. Butler testified that Fuston told him “the guy was getting up or reaching for something, so that's why he shot.” Tr. 1052. The evidence did not support an inference that Fuston did not know “who, if anyone, would be found behind the door,” and that he could not “know[ ] that Mr. Rhodes was on the couch directly inside,” when the uncontroverted evidence showed that Fuston shot because he saw Rhodes “was getting up or reaching for something.” Tr. 1052.

Nor does Butler's testimony that Fuston and the gang were going to the Rhodes home for a “fight” with Dillard support finding that Fuston lacked an intent to kill when he fired after seeing Rhodes move.¹⁰ When asked why he and Fuston had been called to Oklahoma City, Butler said “another fight, I guess.” Tr. 1041.¹¹ Fuston again asks us to speculate that his going to have a “fight” implied an intent to scare Dillard, not an intent to kill Rhodes after Fuston saw him move. But one can fight with an intent to kill the other person, and without more, the “fight” testimony does not compel an inference that Fuston shot Rhodes after seeing him move without intent to kill. See Grant v. Trammell, 727 F.3d 1006, 1014 (10th Cir. 2013).

Fuston's evidence is inadequate compared to evidence that we have found supported a second-degree murder instruction. See Taylor, 554 F.3d at 891 (explaining that the jury could have “believed [the defendant's] testimony and concluded that [he] had no intent to kill [the victim]”); see also Richie, 599 F.3d at 1138 (requiring instruction when forensic evidence suggested the defendant left the victim alive); Phillips, 604 F.3d at 1213–14 (requiring instruction when defendant's statements and witness testimony showed the defendant “may have been severely emotionally disturbed”); Hooks v. Ward, 184 F.3d 1206, 1232 (10th Cir. 1999) (finding instruction warranted because the defendant's statement and post-crime conduct suggested the defendant “merely intended to inflict another one of his ruthless beatings,” not to kill the victim); Turrentine v. Mullin, 390 F.3d 1181, 1193 (10th Cir. 2004) (finding instruction warranted because the defendant presented evidence that he consumed “a combination of anti-depressant medication, anti-anxiety medication, and a large measure of alcohol”).

In sum, the OCCA did not “blunder[ ] so badly that every fairminded jurist would disagree” with its conclusion. Klein v. Martin, 607 U.S. 213, 220–21 (2026). It therefore did not unreasonably apply Beck on this record.

C. Atkins Error

Fuston next argues that the OCCA unreasonably applied Atkins and its progeny. He says Atkins, and in particular Hall v. Florida, 572 U.S. 701 (2014), entitled him to an evidentiary hearing to determine whether he is intellectually disabled. As discussed above, relief under § 2254(d)(1) is available only if the OCCA unreasonably applied clearly established law. Because we conclude that Fuston's entitlement to an Atkins hearing was not clearly established, we affirm the district court's denial of relief.

Oklahoma law makes first-degree murder a capital offense. 21 Okla. Stat. Ann. § 701.9. But it prohibits imposing the death penalty on an intellectually disabled defendant. Id. § 701.10b. A defendant is intellectually disabled if he has “significantly subaverage general intellectual functioning, existing concurrently with significant limitations in adaptive functioning.” Id. § 701.10b(a)(1). And a defendant shows subaverage intellectual functioning by presenting an IQ score of 70 or below, accounting for the standard error of measurement. Id. § 701.10b(c). But “in no event shall a defendant who has received” an IQ score of “seventy-six (76) or above” “be considered intellectually disabled.” Id.

Fuston has four IQ scores relevant here: 59, 69, 75, and 81. Appellant Br. on Ground I at 10. The question is whether Atkins and Hall clearly establish that Oklahoma cannot deny him an Atkins hearing based on his 81 score when he has presented scores below 70.

They do not. Atkins forbids imposing the death penalty on the intellectually disabled as a violation of the Eighth Amendment. 536 U.S. at 316–17. Beyond that general rule, Atkins “le[ft] to the State[s] the task of developing” rules about enforcing this restriction—specifically, the question of “determining which offenders are in fact” intellectually disabled. Id. at 317. Yet Atkins adverted to clinical standards that are used to determine intellectual disability: (1) “subaverage intellectual functioning” and (2) “significant limitations in adaptive skills” that (3) “manifest before age 18.” Id. at 318.

Hall picked up on these clinical requirements. Focusing on subaverage intellectual function, the Court held that “when a defendant's IQ test score falls within the test's acknowledged and inherent margin of error, the defendant must be able to present additional evidence of intellectual disability.” Hall, 572 U.S. at 723. And though the defendant in Hall presented scores between 71 and 80, see id. at 707, the Court did not mention those other scores in its holding—let alone any discussion of how they should affect the analysis. The Court simply noted that “the analysis of multiple IQ scores jointly is a complicated endeavor.” Id. at 714. Hall’s bottom-line holding was this: Florida's law treating an IQ score of 71 as conclusive contravened the Eighth Amendment “[by] failing to take into account the standard error of measurement.” id. at 724.

The Supreme Court's holdings in both Moore v. Texas and Brumfield v. Cain confirm that reading of Hall. Moore explained Hall’s holding: “Hall instructs that, where an IQ score is close to, but above, 70, courts must account for the test's ‘standard error of measurement.’ ” Moore v. Texas, 581 U.S. 1, 13 (2017). The Court then found that because the defendant's IQ score of 74, once adjusted for the SEM, fell “at or below 70,” the state court “had to move on to consider Moore's adaptive functions.” Id. at 14. Brumfield took a similar tack: it concluded that when the defendant presented a score of 75, which fell at 70 when adjusted for the SEM, that score could not conclusively establish that the defendant “could not possess subaverage intelligence.” Brumfield v. Cain, 576 U.S. 305, 316 (2015).

Read together, Hall, Moore, and Brumfield clearly establish that a single IQ score with a SEM reaching below 70 cannot be used as conclusive evidence that the defendant is not intellectually disabled. But Fuston says those cases establish a broader proposition: that a single IQ score with a SEM above 70 cannot conclusively preclude a defendant's intellectual disability when he presents other scores within or below the SEM.

We do not think those cases clearly establish that proposition.¹² The Court recently declined to “provide more detailed guidance beyond what this Court's cases have previously said” about “how courts should assess multiple IQ scores.” Hamm v. Smith (Hamm II), No. 24-872, slip op. at 9 (May 21, 2026) (Sotomayor J., concurring). But what have those cases said about evaluating multiple IQ scores? Apparently, not much. See Hamm v. Smith, 604 U.S. 1, 2 (2024) (per curiam) (noting that neither Hall, Moore, nor Brumfield “specif[y] how courts should evaluate multiple IQ scores”); Hamm II, slip op. at 9 (Sotomayor, J., concurring) (agreeing with decision not to provide, in Hamm, “any meaningful guidance on how courts should assess multiple IQ scores”); id. at 63 (Alito, J., dissenting) (observing the “doctrinal ambiguity and numerous unanswered questions” created by Atkins). So whatever guidance Fuston might wring out of those cases, they little clarify how multiple scores affect the intellectual-disability inquiry. See Busby v. Guerrero, No. 26-70004, 2026 WL 1291044, at *1 (May 8, 2026) (Higginson, J., concurring) (calling for clarity on “whether and how to determine intellectual disability before states may execute defendants”), vacated, No. 25A1235 (U.S. May 14, 2026). They may even undermine Fuston's position. Hamm II, slip op. at 44 (Alito, J., dissenting) (“[S]ufficiently high scores might be dispositive in some cases.” (citing Hall, 572 U.S. at 715)); Brumfield, 576 U.S. at 316 (observing there was no “evidence of any higher IQ test score that could render the state court's determination reasonable”).¹³

Fuston says that Hall’s and Moore’s logic clearly establishes that a score falling above 70 when adjusted for the SEM cannot be treated conclusively. He points out that Hall and Moore involved defendants who presented scores that fell below and above 70 when adjusted for the SEM. So, he says, because the Court still found that the adjusted IQ score falling below 70 entitled the defendant to an Atkins hearing, Hall and Moore clearly establish his entitlement to an Atkins hearing. But the high scores in Hall and Moore played no explicit role in the Court's analysis. Moore, 581 U.S. at 15 (focusing only on the defendant's score of 74 and not considering the effect of his 78 score); Hall, 572 U.S. at 723–24 (focusing on the defendant's score of 71 and ignoring his 80 score). And reasonable jurists disagree on whether Hall and Moore set rules for multiple scores or not. Constrast Hamm II, slip op. at 15 (Sotomayor, J., concurring) (“Thus, in both Hall and Moore, this Court confronted Atkins claimants with multiple scores above 70 (and none below 70), yet still held that courts should consider the scores in light of ‘additional evidence of intellectual disability ․’ ”) (quoting Hall, 572 U.S. at 723), with id. at 46 (Alito, J., dissenting) (“Hall and Moore announced rules about treating a single test score as dispositive ․”).

But even if Hall’s and Moore’s logic supported Fuston, those cases do not clearly establish law. See Andrew, 604 U.S. at 92 (explaining that a habeas petitioner must point to “the holdings, as opposed to the dicta, of th[e] Court's decisions” (quoting White v. Woodall, 572 U.S. 415, 419 (2014))). To be sure, the Court has said that “[g]eneral legal principles” may clearly establish law—but only if “they are holdings of th[e] Court.” Id. at 94. So Fuston would still need to show that Hall or Moore “relie[d] on” the general principle that no score falling above 70 when adjusted for the SEM could be treated as conclusive. Id. at 90. And we do not think the Supreme Court can be said to rely on an unstated, undecided principle. See id. at 92–93 (finding that a general legal principle was a holding in part because the Supreme Court had explicitly stated it).

Fuston also has an alternative argument. He argues that Hall clearly established that the existence of intellectual disability must be determined by reference to clinical standards. And, he says, because clinical standards would not treat the single 81 IQ score as dispositive, neither can the courts. Fuston thus urges a sweeping proposition: “flout[ing] clinical standards” is flouting clearly established law. Appellant Br. on Ground I, at 15. To Fuston, the DSM-5 sets the constitutional standard.

That proposition does not follow from Hall. To be sure, Hall found clinical standards relevant to determining intellectual disability, but Hall does not clearly establish that failing to adopt specific clinical standards to assess intellectual disability violates the Eighth Amendment. See Moore, 581 U.S. at 13 (explaining that though “our precedent” does not “license disregard of current medical standards,” it does not “demand adherence to everything stated in the latest medical guide”); Hall, 572 U.S. at 721 (“[T]he views of medical experts ․ do not dictate the Court's decision, yet the Court does not disregard these informed assessments.”); Smith v. Sharp, 935 F.3d 1064, 1077 (10th Cir. 2019) (explaining that “Atkins clearly establishes that intellectual disability must be assessed, at least in part under the existing clinical definitions applied through expert testimony.” (emphasis added)). The Court's analysis in Hall makes that clear. The Court “consider[ed] the psychiatric and professional studies” informing the use of IQ to determine intellectual disability, Hall, 572 U.S. at 709, and the states’ consensus on using IQ scores, id. at 714 (“A significant majority of States implement the protections of Atkins by taking the SEM into account ․”). And having considered those sources, the Court then “express[ed] its own independent determination reached in light” of clinical standards and state practice. Id. at 710 (emphasis added). Fuston reads Hall as if it ignored state practice and never exercised an independent judgment. Indeed, the cases suggest that a valid IQ score above 70 (as adjusted) standing alone would satisfy the Consitution. Accordingly, we cannot say that Atkins and its progeny clearly establish that Oklahoma's statute violates the Eighth Amendment.

We conclude only that Atkins, Hall, Moore, and Brumfield do not clearly establish that Fuston was entitled to an Atkins hearing. If this case arose in a different posture—without the deference AEDPA requires of us—Atkins and its progeny may well have led us in a different direction. But this case does not present that question.

III. Conclusion

For the foregoing reasons, we affirm the denial of Fuston's habeas petition.

I agree with the majority that Fuston's claim under Beck v. Alabama, 447 U.S. 625 (1980), fails to meet the high bar set by Klein, and therefore I join part II.B. of its opinion. Maj. Op. at 20 (concluding that “the OCCA did not ‘blunder[ ] so badly that every fairminded jurist would disagree’ with its conclusion” regarding the application of Beck) (quoting Klein v. Martin, 607 U.S. 213, 220–21 (2026)). But I disagree with the majority's conclusion that he has failed to make a claim under Atkins v. Virginia, 536 U.S. 304 (2002). I therefore dissent as to part II.C. of the majority's opinion.

The majority finds that the question of whether a defendant with mixed results on IQ tests should receive a hearing is not clearly established, and that therefore Fuston's Atkins claim must fail. Maj. Op. at 3. In my view, the majority asks (and answers) the wrong question.

Plainly, the issue of determining whether a defendant is intellectually disabled when he has multiple IQ scores is unsettled, as demonstrated by the Supreme Court's recent dismissal of the writ of certiorari as improvidently granted in Hamm v. Smith, 146 S. Ct. 1251 (2026) (per curiam). See also id. at 1252 (Sotomayor, J., concurring) (explaining that the Court “should not and cannot use this case to address how courts must analyze multiple IQ scores under Atkins”). But what is not unsettled is when such a person gets a hearing.

In Atkins, the Supreme Court held that the Eighth Amendment prohibits the execution of intellectually disabled criminal defendants. 536 U.S. at 321. An important component of that holding is that a defendant is entitled to a hearing when he “ ‘raise[s] a reasonable doubt as to [his] intellectual disability.’ ” Tryon v. Quick, 81 F.4th 1110, 1143 (10th Cir. 2023) (quoting Brumfield v. Cain, 576 U.S. 305, 313 (2015)). Here, Fuston has made such a showing. He has put forth evidence of multiple IQ results: 59, 69, 75 and 81. ROA Vol. I at 1934. Under Oklahoma law, a defendant is deemed intellectually disabled with a score of 70 or below, but “in no event shall a defendant who has received” a score of 76 or above “be considered intellectually disabled.” Tyron, 81 F.4th at 1144 (quoting Okla. Stat. tit. 21, § 701.10b(C)). He obtained the score of 81 when he was twelve, but there is ample evidence that he could have suffered cognitive decline after that time, including having suffered a head injury while boxing in high school. ROA Vol. I at 201. Given this evidence, he is entitled to an Atkins hearing. Because the district court in this case denied his request for such a hearing, I would reverse and remand the case for a hearing to take place.

The majority essentially holds that Fuston must demonstrate that it is clearly established that he is entitled to a hearing under these circumstances. In my view, this is far too narrow a question given the fact-based nature of the inquiry. He needs to, and has, met the standard set forth in Brumfield, which is clearly established.

The majority's reliance on the Hamm dismissal proves the point. Maj. Op. at 23–24. There, the Court dismissed the writ as improvidently granted because it needed more information. 146 S.Ct. at 1255 (Sotomayor, J., concurring) (noting that “the litigation below did not focus on whether a precise methodology exists that courts must use” in assessing multiple scores, and that “[w]ithout the benefit of [such] an evidentiary record ․ this Court rightly concludes that it should not provide more detailed guidance”). But see id. at 1272 (Alito, J., dissenting) (objecting to the dismissal on the ground that the Court must provide “much-needed guidance to the lower courts” on how to handle multiple scores); id. at 1264 (Thomas, J., dissenting) (stating that he would overrule Atkins on the ground that “[n]othing in the text or history of the Constitution supports [the decision]”). Indeed, a key difference between this case and Hamm is that the defendant there had already received a hearing—it just was incomplete. Id. at 1253–54 (Sotomayor, J., concurring) (summarizing the evidence presented at the hearing). In this case, Fuston has received no such hearing. The message from Hamm is that any decision regarding how to treat multiple IQ scores should not be made in a vacuum. The unsettled question under Supreme Court caselaw, then, is what to do with the evidence once it is presented at a hearing, not whether the hearing should be held in the first place.

Here, the district court denied an Atkins hearing on the ground that it would be moot, as one of Fuston's scores—an 81 he received at age twelve—is above the Oklahoma statutory limit of 76. See Fuston v. Quick, No. CIV-21-179-SLP, 2024 WL 3402535, at *9 (W.D. Okla. July 12, 2024). It is true, as the majority points out, the Court has left it to the states to define what it is to be intellectually disabled. Maj. Op. at 21; Atkins, 536 U.S. at 317 (“[W]e leave to the State[s] the task of developing appropriate ways to enforce the constitutional restriction upon [their] execution of sentences.”). But while the states have broad leeway in determining whether a defendant is intellectually disabled, “[the] States’ discretion ․ is not unfettered.” Moore v. Texas, 581 U.S. 1, 13 (2017); see also id. (“[T]he determination must be informed by the medical community's diagnostic framework.”). Indeed, Hamm itself demonstrates that a state's decision on how to treat intellectual disability under Atkins is subject to review.

Surely, a state cannot treat intellectual disability in a way that contravenes well-established Supreme Court precedent—here, Brumfield—and it was the district court's mistake here to think it could. Accordingly, I would reverse and remand this case for an Atkins hearing.

FOOTNOTES

2.   The statute reads in relevant part:However, in no event shall a defendant who has received an intelligence quotient of seventy-six (76) or above on any individually administered, scientifically recognized, standardized intelligence quotient test administered by a licensed psychiatrist or psychologist, be considered intellectually disabled and, thus, shall not be subject to any proceedings under this section.

3.   The statute provides that:Aggravating circumstances shall be:․2. The defendant knowingly created a great risk of death to more than one person;․7. The existence of a probability that the defendant would commit criminal acts of violence that would constitute a continuing threat to society;․

4.   The parties do not dispute that Fuston's tests yielded a SEM of five points.

5.   Fuston attempts to base his Beck challenge on both § 2254(d)(1) and (d)(2). But Beck challenges, which assess the sufficiency of the evidence—a mixed question of law and fact—are reviewed under § 2254(d)(1). Gilson v. Sirmons, 520 F.3d 1196, 1233–34 (10th Cir. 2008) (concluding that a Beck challenge is “reviewable under § 2254(d)(1)”); see also Grant v. Royal, 886 F.3d 874, 914 (10th Cir. 2018) (noting that § 2254(d)(1) applies to “law-dependent mixed questions”). Section 2254(d)(2), in contrast, applies when the petitioner challenges a state court's determination of a fact question. See Boltz v. Mullin, 415 F.3d 1215, 1230 (10th Cir. 2005).

6.   The state argues that it is not clearly established that Beck applies to discretionary death-penalty statutes, like Oklahoma's. But we found that Beck establishes that proposition in Hooks v. Ward, 184 F.3d 1206, 1227 (10th Cir. 1999) (“Beck entitles a defendant in a capital case to a lesser included instruction when the evidence warrants it, notwithstanding the fact that the jury may retain discretion at sentencing to issue a penalty less than death.”). And several times we have granted relief under § 2254(d)(1) for Beck challenges to convictions under Oklahoma's death-penalty scheme. See, e.g., Richie v. Workman, 599 F.3d 1131, 1141 (10th Cir. 2010) (finding the OCCA's determination “was an unreasonable application of the law clearly established by the Supreme Court in Beck”).

7.   Fuston asks us to ignore the OCCA's finding that no evidence would support second-degree murder because whatever analysis the OCCA said it did, it did not actually do that analysis. But we must “assume the state court is applying the correct federal legal standard when it tells us it is—and when fairly it appears to be—doing just that.” Eizember, 803 F.3d at 1143. Applying the deference AEDPA requires of us, we read its opinion as both telling us that it applied the correct standard and fairly appearing to do so. See Klein v. Martin, 607 U.S. 213, 222 (2026) (explaining that we cannot second-guess whether a state court applied the correct rule when the court “accurately summarized” relevant precedent, “correctly stated the governing rule,” and “stated unequivocally that its decision was based on that rule”).

8.   To the contrary, he encourages us to give “full credence” to Butler's testimony. Appellant Br. on Ground IV at 36.

9.   The court in Taylor found the OCCA's determination was contrary to Beck because it applied the wrong standard, so it reviewed the OCCA determination de novo. See Taylor, 554 F.3d at 886 (“[W]e will analyze the evidence in this case de novo”); id. at 888 (concluding that because the OCCA applied the wrong standard, “we cannot defer to its conclusion that no lesser-included offense instruction was required”).

10.   This fact was not raised below. But even if it were not forfeited, we would still conclude that it did not support finding Fuston shot Rhodes without intent to kill after Rhodes moved.

11.   Fuston misrepresents part of the transcript. He says his statement that “one of the girls” called him and said, “they got into a fight,” means that he went to the Rhodes home to “fight” Dillard. But that is not the import of the testimony; it says only that the girls had fought Dillard before.

12.   We have not squarely addressed whether Atkins and its progeny clearly establish a defendant's entitlement to an Atkins hearing when he presents scores above and within the SEM. See, e.g., Smith v. Sharp, 935 F.3d 1064, 1082–83 (10th Cir. 2019) (addressing scenario where all of defendant's scores were “in the intellectually disabled range”). But we have held that Atkins and Hall clearly establish that the IQ cutoff cannot simply take 70 as a strict cutoff; it must account for the test's SEM. Smith v. Duckworth, 824 F.3d 1233, 1243–44 (10th Cir. 2016).

13.   The dissent contends that we have asked and answered the wrong question in our clearly-established analysis. In the dissent's view, Brumfield clearly establishes Fuston's entitlement to a hearing if he “raise[s] a reasonable doubt as to [his] intellectual disability.” Dissenting Op. 1–2; Brumfield, 576 U.S. at 320. But whether Fuston has raised a reasonable doubt as to his intellectual disability depends on how intellectual disability is determined under Oklahoma law. See Atkins, 536 U.S. at 317 (giving states the responsibility to define intellectual disability). And Oklahoma has determined that a defendant is not intellectually disabled if he presents an IQ score of 76 or above, even if he presents scores below 76. Given that definition of intellectual disability, there could be no reasonable doubt as to Fuston's intellectual disability as Oklahoma defines it. We acknowledge that states do not have “unfettered discretion” to define intellectual disability. Moore, 581 U.S. at 20. But, as we have said, no law clearly establishes that Oklahoma's definition transgresses constitutional bounds. The dissent puts it best: the issue is “[p]lainly ․ unsettled.” Dissenting Op. 1.

TYMKOVICH, Circuit Judge.