Agentic
Like-for-like- GPT-5.5
- 63.9
- Supported · #15/151
- Grok 4.20
- 26.5
- Supported · #144/151
- Basis
- BenchAlign lane · 13 vs 4 public rows
- Reading
- GPT-5.5 leads
GPT-5.5 vs Grok 4.20
Updated September 4, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
Decision reading
GPT-5.5 has the higher public score estimate, 73.27 versus 67.01, but the 90% score intervals overlap. Treat that as a lead, not a settled winner.
14 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
Share or export
Which one for your work
Recommendations appear only when a shared evidence basis or an explicit operating constraint supports the call. Secondary and unsupported use cases stay disclosed below the initial list.
Coding work
Code generation, repair, and software-engineering tasks
GPT-5.5
GPT-5.5 leads on the public coding lane, 67.7 to 28.2, with Supported evidence for both models and non-overlapping 90% intervals.
Confidence: stronger
Agentic work
Tool use, computer use, and multi-step task completion
GPT-5.5
GPT-5.5 leads on the public agentic lane, 63.9 to 26.5, with Supported evidence for both models and non-overlapping 90% intervals.
Confidence: stronger
Long documents
Prompts that approach the documented context limit
Grok 4.20
Grok 4.20 has the larger documented context window.
Confidence: documented
Show secondary and unsupported calls
Chat turn cost
1K fresh input + 500 output tokens
Grok 4.20
Grok 4.20 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Repository review cost
50K fresh input + 3K output tokens
Grok 4.20
Grok 4.20 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
No clear pick
A complete comparable API-rate estimate is not available for both models.
Confidence: rate-fallback
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
24 GPT-5.5 only
14 Shared
9 Grok 4.20 only
- Shared results
- 14
- GPT-5.5 only
- 24
- Grok 4.20 only
- 9
- Like-for-like categories
- 3 / 8
1 category rests on Estimated evidence or different evidence sets. Those rows remain visible for coverage context but do not name a winner.
Category results, on a stated basis
Each row shows the public-lane category score for both models: the BenchAlign lane for agentic, coding, and knowledge, and the weighted public rows elsewhere, with the same rank each model page reports. A row is like-for-like only when both scores rest on Supported evidence or the same weighted set. Directional and not-comparable rows remain visible, but they never receive a winner.
Coding
Like-for-like- GPT-5.5
- 67.7
- Supported · #8/183
- Grok 4.20
- 28.2
- Supported · #171/183
- Basis
- BenchAlign lane · 9 vs 6 public rows
- Reading
- GPT-5.5 leads
Knowledge
Like-for-like- GPT-5.5
- 73.3
- Supported · #7/181
- Grok 4.20
- 48.7
- Supported · #97/181
- Basis
- BenchAlign lane · 6 vs 6 public rows
- Reading
- GPT-5.5 leads · intervals overlap
Multimodal
Directional only- GPT-5.5
- 71.3
- #19/48
- Grok 4.20
- 34.6
- #43/48
- Basis
- Provisional lane · 2 vs 2 weighted rows
- Reading
- Directional only
Reasoning
Not comparable- GPT-5.5
- 63.5
- #15/22
- Grok 4.20
- 34.2
- Unranked · 2 rankable rows
- Basis
- Provisional lane · 2 vs 2 weighted rows
- Reading
- Not comparable
Math
Not comparable- GPT-5.5
- 69.6
- Unranked · 3 rankable rows
- Grok 4.20
- Not ranked
- Basis
- Provisional lane · 2 vs 0 weighted rows
- Reading
- Not comparable
Multilingual
Not comparable- GPT-5.5
- Not ranked
- Grok 4.20
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Instruction following
Not comparable- GPT-5.5
- 92.9
- #7/120
- Grok 4.20
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
| Category | GPT-5.5 | Grok 4.20 | Basis | Reading |
|---|---|---|---|---|
| Agentic | 63.9Supported · #15/151 | 26.5Supported · #144/151 | Like-for-likeBenchAlign lane · 13 vs 4 public rows | GPT-5.5 leads |
| Coding | 67.7Supported · #8/183 | 28.2Supported · #171/183 | Like-for-likeBenchAlign lane · 9 vs 6 public rows | GPT-5.5 leads |
| Knowledge | 73.3Supported · #7/181 | 48.7Supported · #97/181 | Like-for-likeBenchAlign lane · 6 vs 6 public rows | GPT-5.5 leads · intervals overlap |
| Multimodal | 71.3#19/48 | 34.6#43/48 | Directional onlyProvisional lane · 2 vs 2 weighted rows | Directional only |
| Reasoning | 63.5#15/22 | 34.2Unranked · 2 rankable rows | Not comparableProvisional lane · 2 vs 2 weighted rows | Not comparable |
| Math | 69.6Unranked · 3 rankable rows | Not ranked | Not comparableProvisional lane · 2 vs 0 weighted rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Instruction following | 92.9#7/120 | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
Ranks count the models scored in each category’s lane, so the agentic, coding, and knowledge denominators (BenchAlign) differ from the provisional-lane categories. Unranked scores sit on the lane’s scale but never name a winner. Like-for-like rows within 0.5 points read as a practical tie.
Shape of the matched evidence
Only shared public evidence is shown. Sparse evidence stays a ruled list rather than being closed into a radar shape.
Too few matched category axes support a radar. The ruled list below shows only shared benchmark results; positions use each benchmark’s normalized display scale when available.
Terminal-Bench 2.0
Agentic
ARC-AGI-2
Reasoning
HLE w/o tools
Knowledge
SWE-bench Pro
Coding
MMMU-Pro
Multimodal
What each workload costs
Three fixed token mixes turn per-token rates into comparable decisions. Each scenario states context fit and whether cached input had to fall back to the published list-input rate.
Chat turn
1K fresh input + 500 output tokens
- GPT-5.5
- $0.02
- Fits in one request
- Grok 4.20
- $0.005
- Fits in one request
Grok 4.20 has the lower modeled cost
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- GPT-5.5
- $0.34
- Fits in one request
- Grok 4.20
- $0.118
- Fits in one request
Grok 4.20 has the lower modeled cost
Costs use the listed standard API rates.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- GPT-5.5
- $0.5
- Fits in one request
- Grok 4.20
- $0.5
- Fits in one request
- Cached input priced at the published list-input rate
Modeled costs are equal
Grok 4.20 has no published cached-input rate, so cached tokens use its listed input rate.
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationGPT-5.5Grok 4.20
Context window
Maximum documented context; output-token limits may be lower.
GPT-5.5
Grok 4.20
2M
API model ID
GPT-5.5
gpt-5.5
OpenAI pricingGrok 4.20
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
GPT-5.5
$0.5 per 1M cached input tokens
OpenAI pricingGrok 4.20
Not published
Documented inputs
GPT-5.5
Not sourced
Grok 4.20
Not sourced
Documented outputs
GPT-5.5
Not sourced
Grok 4.20
Not sourced
Provider availability
GPT-5.5
Not sourced
Grok 4.20
Not sourced
Reasoning profile
GPT-5.5
Reasoning
Grok 4.20
Reasoning
Weight access
GPT-5.5
Proprietary
Grok 4.20
Proprietary
License
GPT-5.5
Proprietary
Grok 4.20
Proprietary
Release date
GPT-5.5
2026-04-23
Grok 4.20
2026-03-10
If you already use one of these models
- Deployment change
- The models list different providers, so authentication, endpoint behavior, limits, and feature support may change.
- Quality signal
- GPT-5.5 has the higher public score estimate, 73.27 versus 67.01, but the 90% score intervals overlap.
- Workload cost
- Repository review: $0.34 vs $0.118. Cache-heavy agent loop: $0.5 vs $0.5.
- Context tradeoff
- Grok 4.20 has the larger documented window (2M).
Run the same representative tasks against both endpoints before changing production traffic.
Benchmark evidence
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Browse raw public benchmark evidence47 rows
Agentic
Terminal-Bench 2.0
GPT-5.582%SourceGrok 4.2047.1%SourceGPT-5.5 leads this result
CyberGym
GPT-5.581.8%SourceGrok 4.20—Not directly comparable
BrowseComp
GPT-5.584.4%SourceGrok 4.20—Not directly comparable
OSWorld-Verified
GPT-5.578.7%SourceGrok 4.20—Not directly comparable
MCP Atlas
GPT-5.575.3%SourceGrok 4.20—Not directly comparable
Toolathlon
GPT-5.555.6%SourceGrok 4.20—Not directly comparable
τ²-bench results
GPT-5.598%SourceGrok 4.20—Not directly comparable
Gert Labs
Shared sourceGPT-5.572.93%Grok 4.2038.36%GPT-5.5 leads this result
ResearchClawBench
GPT-5.517.0%SourceGrok 4.20—Not directly comparable
OSWorld 2.0
GPT-5.513.0%SourceGrok 4.20—Not directly comparable
JobBench
GPT-5.542.7%SourceGrok 4.20—Not directly comparable
ExploitGym
GPT-5.513.4%SourceGrok 4.20—Not directly comparable
Terminal-Bench 2.1 (Vals)
GPT-5.576.4%SourceGrok 4.2044.2%SourceGPT-5.5 leads this result
DeepSearchQA
GPT-5.5—Grok 4.2062.8%SourceNot directly comparable
Coding
SWE-bench Pro
GPT-5.558.6%SourceGrok 4.2051.8%SourceGPT-5.5 leads this result
Terminal-Bench 2.0
GPT-5.582.0%SourceGrok 4.20—Not directly comparable
Vibe Code Bench
Shared sourceGPT-5.569.85%Grok 4.204.06%GPT-5.5 leads this result
React Native Evals
GPT-5.584.7%SourceGrok 4.20—Not directly comparable
cursorBench31
GPT-5.559.2%SourceGrok 4.20—Not directly comparable
cursorBench32
GPT-5.558.4%SourceGrok 4.20—Not directly comparable
FrontierCode 1.1 Main
GPT-5.543.0%SourceGrok 4.20—Not directly comparable
LiveCodeBench (Vals)
GPT-5.585.3%SourceGrok 4.2084.3%SourceGPT-5.5 leads this result
SWE-bench (Vals)
GPT-5.582.6%SourceGrok 4.2072.2%SourceGPT-5.5 leads this result
LiveCodeBench Pro
GPT-5.5—Grok 4.2074.2%SourceNot directly comparable
SWE-bench Verified
GPT-5.5—Grok 4.2076.7%SourceNot directly comparable
Reasoning
MRCR v2 64K-128K
GPT-5.583.1%SourceGrok 4.20—Not directly comparable
MRCR v2 128K-256K
GPT-5.587.5%SourceGrok 4.20—Not directly comparable
ARC-AGI-2
GPT-5.585%SourceGrok 4.2053.3%SourceGPT-5.5 leads this result
ARC-AGI-3
Shared sourceGPT-5.50.4%Grok 4.200.1%GPT-5.5 leads this result
Knowledge
GPQA
GPT-5.593.6%SourceGrok 4.20—Not directly comparable
GPQA-D
GPT-5.593.6%SourceGrok 4.2088.5%SourceGPT-5.5 leads this result
HLE
GPT-5.552.2%SourceGrok 4.20—Not directly comparable
HLE w/o tools
GPT-5.541.4%SourceGrok 4.2031.6%SourceGPT-5.5 leads this result
GPQA Diamond (Vals)
GPT-5.593.2%SourceGrok 4.2088.6%SourceGPT-5.5 leads this result
MMLU-Pro (Vals)
GPT-5.588.1%SourceGrok 4.2086.3%SourceGPT-5.5 leads this result
HealthBench Hard
GPT-5.5—Grok 4.2020.3%SourceNot directly comparable
MedXpertQA (Text)
GPT-5.5—Grok 4.2050.2%SourceNot directly comparable
Math
Multimodal
MMMU-Pro
GPT-5.581.2%SourceGrok 4.2075.2%SourceGPT-5.5 leads this result
MMMU-Pro w/ Python
GPT-5.583.2%SourceGrok 4.20—Not directly comparable
OfficeQA Pro
GPT-5.554.1%SourceGrok 4.20—Not directly comparable
CharXiv
GPT-5.5—Grok 4.2060.9%SourceNot directly comparable
ERQA
GPT-5.5—Grok 4.2054.1%SourceNot directly comparable
SimpleVQA
GPT-5.5—Grok 4.2057.4%SourceNot directly comparable
MedXpertQA (MM)
GPT-5.5—Grok 4.2065.8%SourceNot directly comparable
Frequently asked questions
Which is better, GPT-5.5 or Grok 4.20?
GPT-5.5 has the higher public score estimate, 73.27 versus 67.01, but the 90% score intervals overlap. The higher estimate is not a decisive winner because the uncertainty ranges overlap.
Which is better for coding, GPT-5.5 or Grok 4.20?
GPT-5.5 leads the public coding lane, 67.7 to 28.2, with Supported evidence for both models and non-overlapping 90% intervals.
Which is better for agentic tasks, GPT-5.5 or Grok 4.20?
GPT-5.5 leads the public agentic tasks lane, 63.9 to 26.5, with Supported evidence for both models and non-overlapping 90% intervals.
Which costs less, GPT-5.5 or Grok 4.20?
For the stated presets, chat costs $0.02 on GPT-5.5 and $0.005 on Grok 4.20; repository review costs $0.34 and $0.118; the cache-heavy agent loop costs $0.5 and $0.5. Grok 4.20 has no published cached-input rate, so cached tokens use its listed input rate.
Which has the larger context window, GPT-5.5 or Grok 4.20?
Grok 4.20 has the larger documented context window: 2M, compared with 1M.
Related comparisons
Last updated September 4, 2026
Watch GPT-5.5 vs Grok 4.20
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.