Noometry

Model comparison

GPT-5.6 Sol vs Grok 4.20 (Non-Reasoning)

GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 48.6 on the Noometry Index. Grok 4.20 (Non-Reasoning) costs 5.1× less per token, which makes it the better buy when GPT-5.6 Sol's lead doesn't matter for your workload.

Last verified . 41 shared benchmarks.

GPT-5.6 Sol OpenAI

65.0

Rank #7 Confirmed

Grok 4.20 (Non-Reasoning) xAI

48.6

Rank #54 Confirmed

Summary

  • They share 41 benchmarks with published results for both. GPT-5.6 Sol scores higher in 9 categories and Grok 4.20 (Non-Reasoning) in 1 category; 8 gaps are clear of the uncertainty.
  • The widest gap is in math, where GPT-5.6 Sol leads 85.6 to 48.2.
  • The biggest single-benchmark swing is ProofBench: 83% for GPT-5.6 Sol and 14% for Grok 4.20 (Non-Reasoning).
  • Grok 4.20 (Non-Reasoning) is cheaper at $1.25 / $2.50 per million input/output tokens, against $4 / $20 for GPT-5.6 Sol.
  • GPT-5.6 Sol accepts more context: 1.05M tokens versus 1M.

Side by side

GPT-5.6 SolGrok 4.20 (Non-Reasoning)
ProviderOpenAIxAI
Noometry Index65.048.6
Released2026-07-092026-02-17
WeightsProprietaryProprietary
Context window1.05M1M
Max output128K30K
Input $ / M tokens$4$1.25
Output $ / M tokens$20$2.50
Results tracked6546

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

Category by category

Coding GPT-5.6 Sol leads

GPT-5.6 Sol: 65.1 (#7), Grok 4.20 (Non-Reasoning): 42.1 (#112)

BenchmarkGPT-5.6 SolGrok 4.20 (Non-Reasoning)
LMArena WebDev16181375
WeirdML89.4%52.3%
LMArena Coding14981459
ALE-Bench2,1771,150
DeepSWE72.7%—
FrontierCode47.5%—
CursorBench41.7%—
FrontierSWE32.2%—
SciCode57.1%—
GSO76.5%—
MirrorCode20%—

Agentic & Tool Use GPT-5.6 Sol leads

GPT-5.6 Sol: 50.3 (#7), Grok 4.20 (Non-Reasoning): 34.4 (#46)

BenchmarkGPT-5.6 SolGrok 4.20 (Non-Reasoning)
τ²-bench Banking46.9%18%
LMArena Search12571189
Vending-Bench 29,6194,663
Terminal-Bench—57.3%
APEX-Agents51.4%—
OSWorld 2.027.3%—
PostTrainBench36.2%—
BALROG60%—
GBAEval52.6%—
GDP.pdf30.7%—

Reasoning GPT-5.6 Sol leads

GPT-5.6 Sol: 74.8 (#8), Grok 4.20 (Non-Reasoning): 52.3 (#32)

BenchmarkGPT-5.6 SolGrok 4.20 (Non-Reasoning)
ARC-AGI-292.5%65.1%
Kagi LLM Benchmark67%75%
NYT Connections (extended)93.8%85.4%
ARC-AGI-197.5%89.5%
Chess Puzzles64%24%
LMArena Hard Prompts14841451
DTBench96%90.1%
LMCA59.2%38.7%
Epoch Capabilities Index161.66151.98
SimpleBench71.7%—
CritPt32.3%—
EnigmaEval37.1%—
Thematic Generalization—63.8%
EBR-Bench44.8%—
Mystery Game Puzzles58%—
Surface Evolver Bench93.1%—
Bench to the Future 30.14—
ForecastBench—61.4

Math GPT-5.6 Sol leads

GPT-5.6 Sol: 85.6 (#9), Grok 4.20 (Non-Reasoning): 48.2 (#65)

BenchmarkGPT-5.6 SolGrok 4.20 (Non-Reasoning)
FrontierMath (Tiers 1-3)89.1%44.9%
FrontierMath Tier 482.9%17.1%
OTIS Mock AIME 2024-2025100%92.2%
ProofBench83%14%
LMArena Math14741455
FrontierMath Erdős0%—

Knowledge GPT-5.6 Sol leads

GPT-5.6 Sol: 64.3 (#18), Grok 4.20 (Non-Reasoning): 52.8 (#60)

BenchmarkGPT-5.6 SolGrok 4.20 (Non-Reasoning)
GPQA Diamond93.5%89.3%
SimpleQA Verified69.7%30.2%
LMArena Expert15161439
Vectara Hallucination Rate12.4%—

Multimodal GPT-5.6 Sol leads

GPT-5.6 Sol: 48.6 (#9), Grok 4.20 (Non-Reasoning): 33.3 (#98)

BenchmarkGPT-5.6 SolGrok 4.20 (Non-Reasoning)
LMArena Vision12811263
Blueprint-Bench 233.6%0%
LMArena Document14831416
Furniture Assembly56.7%—

Multilingual Too close to call

GPT-5.6 Sol: 55.3 (#32), Grok 4.20 (Non-Reasoning): 54.5 (#40)

BenchmarkGPT-5.6 SolGrok 4.20 (Non-Reasoning)
LMArena Non-English14521441
LMArena Chinese15271481
LMArena French14771476
LMArena German14761465
LMArena Japanese14711449
LMArena Korean14421417
LMArena Russian14681458
LMArena Spanish14411443

Instruction Following GPT-5.6 Sol leads

GPT-5.6 Sol: 77.7 (#16), Grok 4.20 (Non-Reasoning): 74.8 (#83)

BenchmarkGPT-5.6 SolGrok 4.20 (Non-Reasoning)
LMArena Instruction Following14821420

Long Context Too close to call

GPT-5.6 Sol: 45.4 (#42), Grok 4.20 (Non-Reasoning): 45.5 (#34)

BenchmarkGPT-5.6 SolGrok 4.20 (Non-Reasoning)
LMArena Longer Query14801437
CL-bench—22.2%
CL-bench Life—11.9%

Writing & Preference GPT-5.6 Sol leads

GPT-5.6 Sol: 73.3 (#12), Grok 4.20 (Non-Reasoning): 65.7 (#44)

BenchmarkGPT-5.6 SolGrok 4.20 (Non-Reasoning)
LMArena Text14571451
LMArena Creative Writing14481438
EQ-Bench Creative Writing19721574
LMArena Multi-Turn14601456
EQ-Bench 41250—

Frequently asked questions

Is GPT-5.6 Sol better than Grok 4.20 (Non-Reasoning)?

GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 48.6 on the Noometry Index. Grok 4.20 (Non-Reasoning) costs 5.1× less per token, which makes it the better buy when GPT-5.6 Sol's lead doesn't matter for your workload.

Which is cheaper, GPT-5.6 Sol or Grok 4.20 (Non-Reasoning)?

Grok 4.20 (Non-Reasoning) is cheaper. It lists at $1.25 per million input tokens and $2.50 per million output tokens; GPT-5.6 Sol lists at $4 and $20.

Is GPT-5.6 Sol or Grok 4.20 (Non-Reasoning) better for coding?

GPT-5.6 Sol scores higher on coding benchmarks: 65.1 versus 42.1 in the Noometry coding category.

Which has the bigger context window?

GPT-5.6 Sol does, with 1.05M tokens against 1M.

How many benchmarks do GPT-5.6 Sol and Grok 4.20 (Non-Reasoning) share?

41 benchmarks have published results for both models. GPT-5.6 Sol has 65 scored results on Noometry and Grok 4.20 (Non-Reasoning) has 46.

Related comparisons

Go deeper