Model comparison

DeepSeek-V3.1-Terminus vs Grok 4.20 (Non-Reasoning)

Grok 4.20 (Non-Reasoning) is the stronger model overall, scoring 48.5 to 42.6 on the Noometry Index. DeepSeek-V3.1-Terminus costs 3.3× less per token, which makes it the better buy when Grok 4.20 (Non-Reasoning)'s lead doesn't matter for your workload.

Last verified October 10, 2026. 14 shared benchmarks.

DeepSeek-V3.1-Terminus

DeepSeek

42.6

Rank #96 Confirmed

Grok 4.20 (Non-Reasoning) xAI

48.5

Rank #48 Confirmed

Summary

They share 14 benchmarks with published results for both. DeepSeek-V3.1-Terminus scores higher in 0 categories and Grok 4.20 (Non-Reasoning) in 7 categories; 5 gaps are clear of the uncertainty.
The widest gap is in reasoning, where Grok 4.20 (Non-Reasoning) leads 52.2 to 25.7.
The biggest single-benchmark swing is Kagi LLM Benchmark: 57.4% for DeepSeek-V3.1-Terminus and 75% for Grok 4.20 (Non-Reasoning).
DeepSeek-V3.1-Terminus is cheaper at $0.30 / $1 per million input/output tokens, against $1.25 / $2.50 for Grok 4.20 (Non-Reasoning).
Grok 4.20 (Non-Reasoning) accepts more context: 1M tokens versus 164K.
DeepSeek-V3.1-Terminus has downloadable open weights; the other is API-only.

Side by side

DeepSeek-V3.1-Terminus and Grok 4.20 (Non-Reasoning) specifications
	DeepSeek-V3.1-Terminus	Grok 4.20 (Non-Reasoning)
Provider	DeepSeek	xAI
Noometry Index	42.6	48.5
Released	2025-09-22	2026-02-17
Weights	Open	Proprietary
Context window	164K	1M
Max output	66K	30K
Input $ / M tokens	$0.30	$1.25
Output $ / M tokens	$1	$2.50
Results tracked	16	46

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

Category by category

Coding Too close to call

DeepSeek-V3.1-Terminus: 41.5 (#114), Grok 4.20 (Non-Reasoning): 41.8 (#104)

Coding benchmarks
Benchmark	DeepSeek-V3.1-Terminus	Grok 4.20 (Non-Reasoning)
LMArena Coding	1426	1458
ALE-Bench	745.17	1,150
LMArena WebDev	—	1374
SciCode	40.6%	—
WeirdML	—	52.3%

Agentic & Tool Use Not comparable

DeepSeek-V3.1-Terminus: —, Grok 4.20 (Non-Reasoning): 34.3 (#43)

Agentic & Tool Use benchmarks
Benchmark	DeepSeek-V3.1-Terminus	Grok 4.20 (Non-Reasoning)
Terminal-Bench	—	57.3%
τ²-bench Banking	—	18%
LMArena Search	—	1189
Vending-Bench 2	—	4,663

Reasoning Grok 4.20 (Non-Reasoning) leads

DeepSeek-V3.1-Terminus: 25.7 (#126), Grok 4.20 (Non-Reasoning): 52.2 (#30)

Reasoning benchmarks
Benchmark	DeepSeek-V3.1-Terminus	Grok 4.20 (Non-Reasoning)
Kagi LLM Benchmark	57.4%	75%
LMArena Hard Prompts	1427	1451
DTBench	81.3%	90.1%
LMCA	28.6%	38.7%
ARC-AGI-2	—	65.1%
NYT Connections (extended)	—	85.4%
ARC-AGI-1	—	89.5%
CritPt	1.7%	—
Chess Puzzles	—	24%
Thematic Generalization	—	63.8%
Epoch Capabilities Index	—	152.01
ForecastBench	—	61.4

Math Grok 4.20 (Non-Reasoning) leads

DeepSeek-V3.1-Terminus: 35.7 (#138), Grok 4.20 (Non-Reasoning): 48.0 (#54)

Math benchmarks
Benchmark	DeepSeek-V3.1-Terminus	Grok 4.20 (Non-Reasoning)
LMArena Math	1402	1455
FrontierMath (Tiers 1-3)	—	44.9%
FrontierMath Tier 4	—	17.1%
OTIS Mock AIME 2024-2025	—	92.2%
ProofBench	—	14%

Knowledge Not comparable

DeepSeek-V3.1-Terminus: —, Grok 4.20 (Non-Reasoning): 52.8 (#54)

Knowledge benchmarks
Benchmark	DeepSeek-V3.1-Terminus	Grok 4.20 (Non-Reasoning)
GPQA Diamond	—	89.3%
SimpleQA Verified	—	30.2%
LMArena Expert	—	1440

Multimodal Not comparable

DeepSeek-V3.1-Terminus: —, Grok 4.20 (Non-Reasoning): 33.2 (#88)

Multimodal benchmarks
Benchmark	DeepSeek-V3.1-Terminus	Grok 4.20 (Non-Reasoning)
LMArena Vision	—	1263
Blueprint-Bench 2	—	0%
LMArena Document	—	1416

Multilingual Grok 4.20 (Non-Reasoning) leads

DeepSeek-V3.1-Terminus: 52.3 (#87), Grok 4.20 (Non-Reasoning): 54.7 (#38)

Multilingual benchmarks
Benchmark	DeepSeek-V3.1-Terminus	Grok 4.20 (Non-Reasoning)
LMArena Non-English	1407	1440
LMArena Russian	1434	1457
LMArena Chinese	—	1481
LMArena French	—	1477
LMArena German	—	1463
LMArena Japanese	—	1450
LMArena Korean	—	1417
LMArena Spanish	—	1442

Instruction Following Too close to call

DeepSeek-V3.1-Terminus: 74.4 (#96), Grok 4.20 (Non-Reasoning): 75.0 (#79)

Instruction Following benchmarks
Benchmark	DeepSeek-V3.1-Terminus	Grok 4.20 (Non-Reasoning)
LMArena Instruction Following	1406	1420

Long Context Grok 4.20 (Non-Reasoning) leads

DeepSeek-V3.1-Terminus: 43.5 (#88), Grok 4.20 (Non-Reasoning): 45.4 (#33)

Long Context benchmarks
Benchmark	DeepSeek-V3.1-Terminus	Grok 4.20 (Non-Reasoning)
LMArena Longer Query	1426	1436
CL-bench	—	22.2%
CL-bench Life	—	11.9%

Writing & Preference Grok 4.20 (Non-Reasoning) leads

DeepSeek-V3.1-Terminus: 61.3 (#87), Grok 4.20 (Non-Reasoning): 65.9 (#42)

Writing & Preference benchmarks
Benchmark	DeepSeek-V3.1-Terminus	Grok 4.20 (Non-Reasoning)
LMArena Text	1420	1451
LMArena Creative Writing	1404	1439
LMArena Multi-Turn	1411	1456
EQ-Bench Creative Writing	—	1574

Frequently asked questions

Is DeepSeek-V3.1-Terminus better than Grok 4.20 (Non-Reasoning)?

Which is cheaper, DeepSeek-V3.1-Terminus or Grok 4.20 (Non-Reasoning)?

DeepSeek-V3.1-Terminus is cheaper. It lists at $0.30 per million input tokens and $1 per million output tokens; Grok 4.20 (Non-Reasoning) lists at $1.25 and $2.50.

Is DeepSeek-V3.1-Terminus or Grok 4.20 (Non-Reasoning) better for coding?

They score almost the same on coding (41.5 vs 41.8); test both on your own repository before choosing.

Which has the bigger context window?

Grok 4.20 (Non-Reasoning) does, with 1M tokens against 164K.

How many benchmarks do DeepSeek-V3.1-Terminus and Grok 4.20 (Non-Reasoning) share?

14 benchmarks have published results for both models. DeepSeek-V3.1-Terminus has 16 scored results on Noometry and Grok 4.20 (Non-Reasoning) has 46.