Model comparison

Nvidia Llama 3.3 Nemotron Super 49b v1.5 vs Step 3.7 Flash

Nvidia Llama 3.3 Nemotron Super 49b v1.5 is the stronger model overall, scoring 40.3 to 37.3 on the Noometry Index.

Last verified October 10, 2026. 0 shared benchmarks.

Nvidia Llama 3.3 Nemotron Super 49b v1.5

NVIDIA

40.3

Rank #151 Confirmed

Step 3.7 Flash

StepFun

37.3

Rank #207 Reported

Summary

The widest gap is in reasoning, where Nvidia Llama 3.3 Nemotron Super 49b v1.5 leads 26.8 to 21.6.
Both cost about the same: $0.40 input and $0.40 output per million tokens.
Step 3.7 Flash accepts more context: 256K tokens versus 131K.

Side by side

Nvidia Llama 3.3 Nemotron Super 49b v1.5 and Step 3.7 Flash specifications
	Nvidia Llama 3.3 Nemotron Super 49b v1.5	Step 3.7 Flash
Provider	NVIDIA	StepFun
Noometry Index	40.3	37.3
Released	2025-07-25	2026-05-29
Weights	Open	Open
Context window	131K	256K
Max output	131K	256K
Input $ / M tokens	$0.40	$0.18
Output $ / M tokens	$0.40	$1.11
Results tracked	12	5

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding Too close to call

Nvidia Llama 3.3 Nemotron Super 49b v1.5: 39.8 (#154), Step 3.7 Flash: 40.0 (#150)

Coding benchmarks
Benchmark	Nvidia Llama 3.3 Nemotron Super 49b v1.5	Step 3.7 Flash
SciCode	—	40%
LMArena Coding	1355	—
ALE-Bench	—	694.12

Reasoning Nvidia Llama 3.3 Nemotron Super 49b v1.5 leads

Nvidia Llama 3.3 Nemotron Super 49b v1.5: 26.8 (#128), Step 3.7 Flash: 21.6 (#219)

Reasoning benchmarks
Benchmark	Nvidia Llama 3.3 Nemotron Super 49b v1.5	Step 3.7 Flash
NYT Connections (extended)	—	39.7%
CritPt	—	2.3%
LMArena Hard Prompts	1336	—

Math Step 3.7 Flash leads

Nvidia Llama 3.3 Nemotron Super 49b v1.5: 38.2 (#141), Step 3.7 Flash: 42.9 (#82)

Math benchmarks
Benchmark	Nvidia Llama 3.3 Nemotron Super 49b v1.5	Step 3.7 Flash
MathArena Final-Answer Competitions	—	68.5%
LMArena Math	1392	—

Knowledge Not comparable

Nvidia Llama 3.3 Nemotron Super 49b v1.5: 36.7 (#165), Step 3.7 Flash: —

Knowledge benchmarks
Benchmark	Nvidia Llama 3.3 Nemotron Super 49b v1.5	Step 3.7 Flash
LMArena Expert	1330	—

Multilingual Not comparable

Nvidia Llama 3.3 Nemotron Super 49b v1.5: 45.5 (#168), Step 3.7 Flash: —

Multilingual benchmarks
Benchmark	Nvidia Llama 3.3 Nemotron Super 49b v1.5	Step 3.7 Flash
LMArena Non-English	1316	—
LMArena Japanese	1300	—
LMArena Russian	1332	—

Instruction Following Not comparable

Nvidia Llama 3.3 Nemotron Super 49b v1.5: 68.6 (#188), Step 3.7 Flash: —

Instruction Following benchmarks
Benchmark	Nvidia Llama 3.3 Nemotron Super 49b v1.5	Step 3.7 Flash
LMArena Instruction Following	1299	—

Long Context Not comparable

Nvidia Llama 3.3 Nemotron Super 49b v1.5: 40.0 (#164), Step 3.7 Flash: —

Long Context benchmarks
Benchmark	Nvidia Llama 3.3 Nemotron Super 49b v1.5	Step 3.7 Flash
LMArena Longer Query	1315	—

Writing & Preference Not comparable

Nvidia Llama 3.3 Nemotron Super 49b v1.5: 53.1 (#159), Step 3.7 Flash: —

Writing & Preference benchmarks
Benchmark	Nvidia Llama 3.3 Nemotron Super 49b v1.5	Step 3.7 Flash
LMArena Text	1338	—
LMArena Creative Writing	1307	—
LMArena Multi-Turn	1334	—

Frequently asked questions

Is Nvidia Llama 3.3 Nemotron Super 49b v1.5 better than Step 3.7 Flash?

Nvidia Llama 3.3 Nemotron Super 49b v1.5 is the stronger model overall, scoring 40.3 to 37.3 on the Noometry Index.

Which is cheaper, Nvidia Llama 3.3 Nemotron Super 49b v1.5 or Step 3.7 Flash?

Nvidia Llama 3.3 Nemotron Super 49b v1.5 is cheaper. It lists at $0.40 per million input tokens and $0.40 per million output tokens; Step 3.7 Flash lists at $0.18 and $1.11.

Is Nvidia Llama 3.3 Nemotron Super 49b v1.5 or Step 3.7 Flash better for coding?

They score almost the same on coding (39.8 vs 40.0); test both on your own repository before choosing.

Which has the bigger context window?

Step 3.7 Flash does, with 256K tokens against 131K.

How many benchmarks do Nvidia Llama 3.3 Nemotron Super 49b v1.5 and Step 3.7 Flash share?

0 benchmarks have published results for both models. Nvidia Llama 3.3 Nemotron Super 49b v1.5 has 12 scored results on Noometry and Step 3.7 Flash has 5.