Beam 1.0

Beam 1.0 ranks first on LongBench v2.

Reflection Released 5 Oct 2026

Benchmarks with a score

Ranked first, out of five or more models

In the top three

Reflection Oct 2026

Open weights

150.2 #60 of 271

Epoch Capabilities Index

Where it ranks

Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.

Benchmark	Score	Rank	Source
LongBench v2	66.9%	#1 of 27	Model card
AA-LCR (Artificial Analysis Long Context Reasoning)	79.3%	#2 of 31	Model card
IFBench	79.7%	#10 of 70	Model card
Humanity's Last Exam	51.2%	#33 of 147	Model card
LMArena (Elo)	1,436	#82 of 333	Official leaderboard

Benchmark	Score	Rank	Source
GPQA Diamond	90.5%	#25 of 284	Model card
AIME 2026	97.8%	#4 of 34	Model card
CritPt	16.3%	#7 of 14	Model card
ARC-AGI-2	23.5%	#49 of 95	Official leaderboard
ARC-AGI	74%	#50 of 92	Official leaderboard
ProofBench	18%	#56 of 83	Epoch AI
MysteryMechanism	22.1%	#21 of 29	Official leaderboard
ProofBench v1.1	18%	#42 of 51	Official leaderboard

Benchmark	Score	Rank	Source
SWE-bench Verified	80.9%	#9 of 150	Model card
SWE-Bench Pro	65.5%	#12 of 77	Model card
SciCode	49.7%	#10 of 47	Model card
SWE-bench Multilingual	78%	#19 of 70	Model card
Vibe Code Bench	66.4%	#49 of 115	Official leaderboard
Code Arena	1,436	#70 of 124	Official leaderboard
Code Migration v1.1	22.9%	#25 of 29	Official leaderboard
DeepSWE	44.4%	#46 of 51	Model card
IOI v2 (Vals)	22.8%	#41 of 45	Official leaderboard
FrontierCode 2.0	2.90%	#25 of 26	Official leaderboard
SWE Atlas	34.6%	#11 of 11	Model card
SWE-Bench Pro V2 Hard	77.2%	#1 of 1	Model card

Benchmark	Score	Rank	Source
τ³-bench Banking	38%	#4 of 20	Model card
MCP Atlas	78.7%	#14 of 40	Model card
BrowseComp	77.4%	#34 of 77	Model card
AutomationBench	37%	#20 of 41	Model card
Terminal-Bench 2.1	80.1%	#26 of 53	Model card
DeepSearchQA	80.1%	#14 of 22	Model card
Agent Arena	-0.05	#39 of 53	Official leaderboard

Benchmark	Score	Rank	Source
APEX-Agents	48.8%	#32 of 56	Epoch AI
Legal Research Bench	28.9%	#55 of 88	Official leaderboard
Harvey LAB v1.1 (HLAB-Vals-1.1)	4.58%	#27 of 33	Official leaderboard
Excel Modeling Benchmark v1.1	64.7%	#31 of 37	Official leaderboard
Tax Agent Bench	8.81%	#67 of 79	Official leaderboard
Vals Legal Index	61.8%	#29 of 33	Official leaderboard
Vals Index v2.2	43.4%	#29 of 32	Official leaderboard
Vals Finance Agent v2.1	51.8%	#32 of 35	Official leaderboard
Vals Finance Index	46.5%	#32 of 35	Official leaderboard

Benchmark	Score	Rank	Source
Human Pathogen Capabilities Test v2 (HPCT v2)	59.6%	#30 of 71	Official leaderboard
Agentic BAIM-LLM Evaluation (ABLE)	50.5%	#9 of 16	Official leaderboard
TxBench-Antibody-Discovery	34%	#20 of 28	Official leaderboard
TxBench-Oligo-Discovery	27.4%	#24 of 28	Official leaderboard
TxBench-Preclinical-Pharmacology	39.7%	#24 of 28	Official leaderboard
BioSecBench-Refusal	3.40%	#19 of 22	Official leaderboard
OncologyBench	1.33%	#29 of 30	Official leaderboard

Benchmark	Score		Rank	Source
Vals Science Index	22%		#23 of 23	Official leaderboard

Reflection's models in color, against every other model on the Epoch Capabilities Index, by release date.

The benchmarks the lab chose to headline, with the scores it gave.