Enter opens the highlighted result. Esc closes.

GPT-6 Sol vs Hy4 flash

GPT-6 Sol leads Hy4 flash on coding, 72.2 to 44.5 on the Vibe Coding Index, and Hy4 flash costs 71.2x less per output token ($0.1405 vs $10.00 per million). Updated November 22, 2026; 10 benchmarks have a result for both.

GPT-6 SolOpenAI
Coding
72.2
Overall
73.7
Output $/M
$10.00
8
Hy4 flashTencent
Coding
44.5
Overall
No data
Output $/M
$0.1405
2
vs boards won of 10 shared

GPT-6 Sol or Hy4 flash: which to pick

Pick GPT-6 Sol for coding: it scores 72.2 on the Vibe Coding Index against 44.5 for Hy4 flash.

Pick Hy4 flash for output tokens 71.2x cheaper ($0.1405 vs $10.00 per million) and work measured by Tax Agent Bench (Vals AI) or Legal Research Bench (Vals AI), where it leads.

Shared results
10
GPT-6 Sol only
101
Hy4 flash only
2
Like-for-like categories
3 of 9

Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.

Which one for your work

Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.

At a glance

Coding index
GPT-6 Sol 72.2 from 4 benchmarks, Hy4 flash 44.5 from 2 benchmarks.
Widest category
Coding, 26.8 points apart on average over 3 shared benchmarks.
Widest single gap
Terminal-Bench 4 (Vals AI): GPT-6 Sol 44.4% against Hy4 flash 15.2%, 39.1 points apart on a 0 to 100 scale.
Output price
GPT-6 Sol $10.00 against Hy4 flash $0.1405 per million output tokens, 71.2x apart.
Context window
GPT-6 Sol takes 1.05M tokens, Hy4 flash 1.05M.

Benchmark by benchmark

First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.

Boards won, where both have a result
CategoryGPT-6 SolBarsHy4 flashTies
Coding30
Agentic work02
Web design50
All shared boards82
CodingLike-for-likeSix of these boards feed the Vibe Coding Index30
Coding benchmarks for GPT-6 Sol and Hy4 flash
BenchmarkGPT-6 SolHy4 flashLeader
LMArena WebDev1674Not measuredOnly GPT-6 Sol measured
Terminal-Bench 4 (Vals AI)44.4%15.2%GPT-6 Sol
Vibe Code Bench (Vals AI)87.8%74.8%GPT-6 Sol
Code migration (Vals AI)57.2%Not measuredOnly GPT-6 Sol measured
ProgramBench (Vals AI)2.0%Not measuredOnly GPT-6 Sol measured
IOI (Vals AI)82.6%Not measuredOnly GPT-6 Sol measured
Terminal-Bench 2.1 (Vals AI)83.2%Not measuredOnly GPT-6 Sol measured
CyberBench (Vals AI)78.0%Not measuredOnly GPT-6 Sol measured
SciCode57.6%Not measuredOnly GPT-6 Sol measured
Terminal-Bench 4.0 (Artificial Analysis run)43.9%Not measuredOnly GPT-6 Sol measured
FrontierCode 2.035.6%Not measuredOnly GPT-6 Sol measured
FrontierCode49.3%Not measuredOnly GPT-6 Sol measured
MirrorCode48.9%Not measuredOnly GPT-6 Sol measured
ALE-Bench2462.0Not measuredOnly GPT-6 Sol measured
LMArena Text: coding1448Not measuredOnly GPT-6 Sol measured
LMArena Text: software and IT1433Not measuredOnly GPT-6 Sol measured
LMArena WebDev: React1676Not measuredOnly GPT-6 Sol measured
LMArena WebDev: HTML1684Not measuredOnly GPT-6 Sol measured
LMArena WebDev: from an image1659Not measuredOnly GPT-6 Sol measured
GameBench 279.050.5GPT-6 Sol
Terminal-Bench 4.049.4%Not measuredOnly GPT-6 Sol measured
Neither model has a result on the other 31 boards here
IntelligenceDirectional onlyPublic boards, shown as published, kept out of the index2 boards, none shared
Intelligence benchmarks for GPT-6 Sol and Hy4 flash
BenchmarkGPT-6 SolHy4 flashLeader
AA Intelligence Index44.2Not measuredOnly GPT-6 Sol measured
Epoch Capabilities Index163.8Not measuredOnly GPT-6 Sol measured
Agentic workLike-for-likePublic boards, shown as published, kept out of the index02
Agentic work benchmarks for GPT-6 Sol and Hy4 flash
BenchmarkGPT-6 SolHy4 flashLeader
Design Arena: agents building web apps1201Not measuredOnly GPT-6 Sol measured
Design Arena: agents building full-stack apps1210Not measuredOnly GPT-6 Sol measured
Design Arena: agents building mobile apps1192Not measuredOnly GPT-6 Sol measured
Design Arena: agents building Android apps1169Not measuredOnly GPT-6 Sol measured
Design Arena: spreadsheets937Not measuredOnly GPT-6 Sol measured
SREBench (Vals AI)30.5%Not measuredOnly GPT-6 Sol measured
Finance Agent (Vals AI)49.0%Not measuredOnly GPT-6 Sol measured
Tax Agent Bench (Vals AI)15.0%23.3%Hy4 flash
Legal Research Bench (Vals AI)28.9%38.5%Hy4 flash
Harvey's Legal Agent Benchmark (Vals AI)Not measured5.4%Only Hy4 flash measured
Excel Modeling Benchmark (Vals AI)71.5%Not measuredOnly GPT-6 Sol measured
BioMysteryBench (Vals AI)74.8%Not measuredOnly GPT-6 Sol measured
Terminal-Bench-Science (Vals AI)30.0%Not measuredOnly GPT-6 Sol measured
Vals Science Index45.0%Not measuredOnly GPT-6 Sol measured
APEX-Agents54.3%Not measuredOnly GPT-6 Sol measured
SpireClimb9.2%Not measuredOnly GPT-6 Sol measured
Vending-Bench 2$14,428Not measuredOnly GPT-6 Sol measured
LMArena Search1174Not measuredOnly GPT-6 Sol measured
Neither model has a result on the other 26 boards here
ReasoningDirectional onlyPublic boards, shown as published, kept out of the index18 boards, none shared
Reasoning benchmarks for GPT-6 Sol and Hy4 flash
BenchmarkGPT-6 SolHy4 flashLeader
MysteryMechanism (Vals AI)30.2%Not measuredOnly GPT-6 Sol measured
Humanity's Last Exam (Artificial Analysis run)47.9%Not measuredOnly GPT-6 Sol measured
HLE-Diamond (Artificial Analysis run)45.2%Not measuredOnly GPT-6 Sol measured
CritPt v2, hard subset (Artificial Analysis run)66.5%Not measuredOnly GPT-6 Sol measured
GPQA Diamond (Epoch AI run)94.3%Not measuredOnly GPT-6 Sol measured
ARC-AGI95.5%Not measuredOnly GPT-6 Sol measured
ARC-AGI-289.6%Not measuredOnly GPT-6 Sol measured
SimpleBench73.1%Not measuredOnly GPT-6 Sol measured
Chess puzzles56.0%Not measuredOnly GPT-6 Sol measured
Mystery game puzzles56.0%Not measuredOnly GPT-6 Sol measured
EBR-bench53.3%Not measuredOnly GPT-6 Sol measured
LMArena Text: hard prompts1421Not measuredOnly GPT-6 Sol measured
LMArena Text: expert1439Not measuredOnly GPT-6 Sol measured
LMArena Text: hard prompts in English1417Not measuredOnly GPT-6 Sol measured
LMArena Vision: diagrams1263Not measuredOnly GPT-6 Sol measured
LMArena Vision: homework1290Not measuredOnly GPT-6 Sol measured
Humanity's Last Exam (Scale run)33.8%Not measuredOnly GPT-6 Sol measured
GoBench2003Not measuredOnly GPT-6 Sol measured
Neither model has a result on the other 8 boards here
MathDirectional onlyPublic boards, shown as published, kept out of the index7 boards, none shared
Math benchmarks for GPT-6 Sol and Hy4 flash
BenchmarkGPT-6 SolHy4 flashLeader
OTIS Mock AIME100.0%Not measuredOnly GPT-6 Sol measured
FrontierMath, tiers 1 to 389.8%Not measuredOnly GPT-6 Sol measured
FrontierMath, tier 490.0%Not measuredOnly GPT-6 Sol measured
ProofBench83.0%Not measuredOnly GPT-6 Sol measured
LMArena Text: math1400Not measuredOnly GPT-6 Sol measured
LMArena Text: mathematical work1413Not measuredOnly GPT-6 Sol measured
MathArena76.2%Not measuredOnly GPT-6 Sol measured
Neither model has a result on the other 7 boards here
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
Knowledge benchmarks for GPT-6 Sol and Hy4 flash
BenchmarkGPT-6 SolHy4 flashLeader
MedCode (Vals AI)47.1%Not measuredOnly GPT-6 Sol measured
AA-Omniscience21.4Not measuredOnly GPT-6 Sol measured
SimpleQA Verified60.7%Not measuredOnly GPT-6 Sol measured
LMArena Text: science1388Not measuredOnly GPT-6 Sol measured
LMArena Text: medicine and healthcare1362Not measuredOnly GPT-6 Sol measured
LMArena Text: legal and government1400Not measuredOnly GPT-6 Sol measured
LMArena Text: business and finance1384Not measuredOnly GPT-6 Sol measured
FACTS Benchmark Suite74.0%Not measuredOnly GPT-6 Sol measured
Neither model has a result on the other 13 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index50
Web design benchmarks for GPT-6 Sol and Hy4 flash
BenchmarkGPT-6 SolHy4 flashLeader
Design Arena: websites12841227GPT-6 Sol
Design Arena: UI components12811262GPT-6 Sol
Design Arena: game dev12891202GPT-6 Sol
Design Arena: data viz1271Not measuredOnly GPT-6 Sol measured
Design Arena: 3D12891209GPT-6 Sol
Design Arena: all code categories12871225GPT-6 Sol
Design Arena: SVG1296Not measuredOnly GPT-6 Sol measured
Design Arena: ASCII art1212Not measuredOnly GPT-6 Sol measured
Neither model has a result on the other 1 board here
GeneralDirectional onlyPublic boards, shown as published, kept out of the index31 boards, none shared
General benchmarks for GPT-6 Sol and Hy4 flash
BenchmarkGPT-6 SolHy4 flashLeader
Vals IndexNot measured50.8%Only Hy4 flash measured
Public Benefits Bench (Vals AI)56.6%Not measuredOnly GPT-6 Sol measured
MedScribe (Vals AI)82.0%Not measuredOnly GPT-6 Sol measured
Long context reasoning (AA-LCR)83.7%Not measuredOnly GPT-6 Sol measured
GDPval-AA1514Not measuredOnly GPT-6 Sol measured
DTBench97.3%Not measuredOnly GPT-6 Sol measured
LMCA59.1%Not measuredOnly GPT-6 Sol measured
Furniture assembly58.3%Not measuredOnly GPT-6 Sol measured
LMArena Text1398Not measuredOnly GPT-6 Sol measured
LMArena Text: instruction following1411Not measuredOnly GPT-6 Sol measured
LMArena Text: creative writing1371Not measuredOnly GPT-6 Sol measured
LMArena Vision1232Not measuredOnly GPT-6 Sol measured
LMArena Text: English1398Not measuredOnly GPT-6 Sol measured
LMArena Text: non-English1387Not measuredOnly GPT-6 Sol measured
LMArena Text: Chinese1408Not measuredOnly GPT-6 Sol measured
LMArena Text: German1388Not measuredOnly GPT-6 Sol measured
LMArena Text: Japanese1358Not measuredOnly GPT-6 Sol measured
LMArena Text: Korean1342Not measuredOnly GPT-6 Sol measured
LMArena Text: French1400Not measuredOnly GPT-6 Sol measured
LMArena Text: Spanish1394Not measuredOnly GPT-6 Sol measured
LMArena Text: Russian1407Not measuredOnly GPT-6 Sol measured
LMArena Text: Polish1388Not measuredOnly GPT-6 Sol measured
LMArena Text: multi-turn1413Not measuredOnly GPT-6 Sol measured
LMArena Text: longer queries1415Not measuredOnly GPT-6 Sol measured
LMArena Text: writing and language1390Not measuredOnly GPT-6 Sol measured
LMArena Text: entertainment and media1369Not measuredOnly GPT-6 Sol measured
LMArena Vision: English1227Not measuredOnly GPT-6 Sol measured
LMArena Vision: Chinese1268Not measuredOnly GPT-6 Sol measured
LMArena Vision: reading text1251Not measuredOnly GPT-6 Sol measured
LMArena Vision: humor1173Not measuredOnly GPT-6 Sol measured
LMArena Vision: creative writing1190Not measuredOnly GPT-6 Sol measured
Neither model has a result on the other 18 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for GPT-6 Sol and Hy4 flash
BenchmarkGPT-6 SolHy4 flashLeader
Neither model has a result on 5 boards here

What each workload costs

Three fixed token mixes at each model's listed price, so per-token rates become a bill.

Chat turn

1K input + 500 output tokens

GPT-6 Sol$0.0070Fits in one request
Hy4 flash$0.0001Fits in one request

Hy4 flash costs less

Repository review

50K input + 3K output tokens

GPT-6 Sol$0.130Fits in one request
Hy4 flash$0.0022Fits in one request

Hy4 flash costs less

Agent loop

200K input + 10K output tokens

GPT-6 Sol$0.500Fits in one request
Hy4 flash$0.0084Fits in one request

Hy4 flash costs less

Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.

Specification differences

The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.

Specifications of GPT-6 Sol and Hy4 flash
SpecificationOpenAI GPT-6 SolTencent Hy4 flash
MakerOpenAITencent
WeightsNo public weights listedOpen weights
ReleasedSeptember 22, 2026November 16, 2026
Context window1.05M1.05M
Input $/Mtok$2.00$0.035
Output $/Mtok$10.00$0.1405
Coding index72.244.5
Independent measurements behind it42
Cost per solved taskNot computableNot computable
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-6 Sol better than Hy4 flash for coding?

Pick GPT-6 Sol for coding: it scores 72.2 on the Vibe Coding Index against 44.5 for Hy4 flash. Pick Hy4 flash for output tokens 71.2x cheaper ($0.1405 vs $10.00 per million) and work measured by Tax Agent Bench (Vals AI) or Legal Research Bench (Vals AI), where it leads.

Which is cheaper, GPT-6 Sol or Hy4 flash?

GPT-6 Sol costs $2.00 per million input tokens and $10.00 per million output tokens; Hy4 flash costs $0.035 and $0.1405. Cost per solved task is known for neither of them.

Which has the bigger context window, GPT-6 Sol or Hy4 flash?

GPT-6 Sol, with 1.05M tokens against 1.05M.

How many benchmarks is the GPT-6 Sol vs Hy4 flash comparison based on?

10 benchmarks have a result for both models as of November 22, 2026. GPT-6 Sol's coding index rests on 4 independent benchmarks and Hy4 flash's on 2. Where only one model has a result, the page says so instead of naming a winner.

Compare next

Embed this comparison on your site

Free under CC BY 4.0, while the source link stays visible. Preview the frame · Terms and other embeds

<iframe src="https://llmbenchmark.ai/embed/compare/gpt-6-sol-vs-tencent-hy4-flash" title="GPT-6 Sol vs Hy4 flash, LLMBenchmark.ai" width="100%" height="360" style="border:0;max-width:100%" loading="lazy"></iframe>
<p style="font:13px/1.4 sans-serif;margin:6px 0 0">Source: <a href="https://llmbenchmark.ai/compare/gpt-6-sol-vs-tencent-hy4-flash">LLMBenchmark.ai, GPT-6 Sol vs Hy4 flash</a></p>