Coding work
Code generation, repair, and software-engineering tasks
GPT-5.4
GPT-5.4 leads on the public coding lane, 49.7 to 21, with Supported evidence for both models and non-overlapping 90% intervals.
Confidence: stronger
GPT-5.4 vs Trinity-Large-Thinking
Updated September 29, 2026. Rank says GPT-5.4 is ahead. Price, access, and your workload can each overturn that. Public scores include evidence status and uncertainty.
Both of these models will change. Get the price, version and retirement notices for the pair, each with its source. Follow model changes
Share or export
Decision reading
GPT-5.4 has the higher public score, 68.44 versus 34.23, and the 90% score intervals do not overlap. 2 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
- Shared results
- 2
- GPT-5.4 only
- 36
- Trinity-Large-Thinking only
- 3
- Like-for-like categories
- 1 / 8
Supported: GPT-5.4 · Estimated: Trinity-Large-ThinkingHow the comparison works
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.
Long documents
Prompts that approach the documented context limit
GPT-5.4
GPT-5.4 has the larger documented context window.
Confidence: documentedChat turn cost
1K fresh input + 500 output tokens
Trinity-Large-Thinking
Trinity-Large-Thinking has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Show secondary and unsupported calls
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Trinity-Large-Thinking
Trinity-Large-Thinking has the lower estimated token cost for this stated workload. Trinity-Large-Thinking has no published cached-input rate, so cached tokens use its listed input rate.
Confidence: rate-fallbackRepository review cost
50K fresh input + 3K output tokens
Trinity-Large-Thinking
Trinity-Large-Thinking has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-ratesAgentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
Trinity-Large-Thinking is scored on Estimated evidence for agentic, so the reading is directional rather than like-for-like.
Confidence: limited
Which one for a specific job
Choose a job from the LLM Selector's task catalog to see the category row it rests on, under the same basis rules as the table below. A directional row stays directional; choosing a job never creates a winner.
The same task catalog as the LLM Selector. Each job names the evidence surface it rests on; nothing here adds a new score.
49.7GPT-5.421.0Trinity-Large-Thinking
Like-for-like · BenchAlign v5.7
GPT-5.4 leads the like-for-like coding row.
Coding scores combine specific tasks and setups. Match the editor, harness, and effort to your workflow.
Same basis rules as the category table below
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
36 GPT-5.4 only
2 Shared
3 Trinity-Large-Thinking only
3 categories rest on Estimated evidence or different evidence sets. Those rows remain visible for coverage context but do not name a winner.
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.
A shared-evidence shape is not available.
BenchLM does not draw a radar or infer missing axes when the matched evidence is too sparse.
Category results, on a stated basis
Each row shows the public-lane category score for both models: the BenchAlign v5.7 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-5.4
- 49.7
- Supported · #40/142
- Trinity-Large-Thinking
- 21.0
- Supported · #119/142
- Basis
- BenchAlign v5.7 lane · 4 vs 1 public rows
- Reading
- GPT-5.4 leads
Agentic
Directional only- GPT-5.4
- 47.8
- Supported · #41/117
- Trinity-Large-Thinking
- 17.7
- Estimated · #104/117
- Basis
- BenchAlign v5.7 lane · 14 vs 1 public rows
- Reading
- Directional only
Knowledge
Directional only- GPT-5.4
- 66.7
- Supported · #19/168
- Trinity-Large-Thinking
- 35.9
- Estimated · #114/168
- Basis
- BenchAlign v5.7 lane · 7 vs 2 public rows
- Reading
- Directional only
Instruction following
Directional only- GPT-5.4
- 89.2
- #19/124
- Trinity-Large-Thinking
- 66.3
- #68/124
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Directional only
Reasoning
Not comparable- GPT-5.4
- 60.6
- #21/27
- Trinity-Large-Thinking
- 48.3
- Unranked · 2 rankable rows
- Basis
- Provisional lane · 2 vs 0 weighted rows
- Reading
- Not comparable
Multimodal
Not comparable- GPT-5.4
- 69.3
- #21/50
- Trinity-Large-Thinking
- Not ranked
- Basis
- Provisional lane · 3 vs 0 weighted rows
- Reading
- Not comparable
Multilingual
Not comparable- GPT-5.4
- Not ranked
- Trinity-Large-Thinking
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Math
Not comparable- GPT-5.4
- 64.4
- Unranked · 2 rankable rows
- Trinity-Large-Thinking
- Not ranked
- Basis
- Provisional lane · 2 vs 0 weighted rows
- Reading
- Not comparable
| Category | GPT-5.4 | Trinity-Large-Thinking | Basis | Reading |
|---|---|---|---|---|
| Coding | 49.7Supported · #40/142 | 21.0Supported · #119/142 | Like-for-likeBenchAlign v5.7 lane · 4 vs 1 public rows | GPT-5.4 leads |
| Agentic | 47.8Supported · #41/117 | 17.7Estimated · #104/117 | Directional onlyBenchAlign v5.7 lane · 14 vs 1 public rows | Directional only |
| Knowledge | 66.7Supported · #19/168 | 35.9Estimated · #114/168 | Directional onlyBenchAlign v5.7 lane · 7 vs 2 public rows | Directional only |
| Instruction following | 89.2#19/124 | 66.3#68/124 | Directional onlyProvisional lane · 0 vs 0 weighted rows | Directional only |
| Reasoning | 60.6#21/27 | 48.3Unranked · 2 rankable rows | Not comparableProvisional lane · 2 vs 0 weighted rows | Not comparable |
| Multimodal | 69.3#21/50 | Not ranked | Not comparableProvisional lane · 3 vs 0 weighted rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Math | 64.4Unranked · 2 rankable rows | Not ranked | Not comparableProvisional lane · 2 vs 0 weighted rows | Not comparable |
Ranks count the models scored in each category’s lane, so the agentic, coding, and knowledge denominators (BenchAlign v5.7) differ from the provisional-lane categories. Unranked scores sit on the provisional lane’s scale but never name a winner. Like-for-like rows within 0.5 points read as a practical tie.
Supported evidence per lane · bars run 0–100Methodology
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-5.4
- $0.01
- Fits in one request
- Trinity-Large-Thinking
- $0.0007
- Fits in one request
Trinity-Large-Thinking has the lower modeled cost
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- GPT-5.4
- $0.17
- Fits in one request
- Trinity-Large-Thinking
- $0.0152
- Fits in one request
Trinity-Large-Thinking 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-5.4
- $0.25
- Fits in one request
- Trinity-Large-Thinking
- $0.064
- Fits in one request
- Cached input priced at the published list-input rate
Trinity-Large-Thinking has the lower modeled cost
Trinity-Large-Thinking has no published cached-input rate, so cached tokens use its listed input rate.
Cached input falls back to the list input rate only where a cached rate is unpublished
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationGPT-5.4Trinity-Large-Thinking
Context window
Maximum documented context; output-token limits may be lower.
GPT-5.4
1.05M
OpenAI pricingTrinity-Large-Thinking
512K
API model ID
GPT-5.4
gpt-5.4
OpenAI pricingTrinity-Large-Thinking
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
GPT-5.4
$0.25 per 1M cached input tokens
OpenAI pricingTrinity-Large-Thinking
Not published
Documented inputs
GPT-5.4
Not sourced
Trinity-Large-Thinking
Not sourced
Documented outputs
GPT-5.4
Not sourced
Trinity-Large-Thinking
Not sourced
Provider availability
GPT-5.4
Not sourced
Trinity-Large-Thinking
Not sourced
Reasoning profile
GPT-5.4
Reasoning
Trinity-Large-Thinking
Reasoning
Weight access
GPT-5.4
Proprietary
Trinity-Large-Thinking
Open Weight
License
GPT-5.4
Proprietary
Trinity-Large-Thinking
Open Weight
Release date
GPT-5.4
2026-03-05
Trinity-Large-Thinking
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-5.4 has the higher public score, 68.44 versus 34.23, and the 90% score intervals do not overlap.
- Workload cost
- Repository review: $0.17 vs $0.0152. Cache-heavy agent loop: $0.25 vs $0.064.
- Context tradeoff
- GPT-5.4 has the larger documented window (1.05M).
Run the same representative tasks against both endpoints before changing production traffic.
Questions
Which is better, GPT-5.4 or Trinity-Large-Thinking?
GPT-5.4 has the higher public score, 68.44 versus 34.23, and the 90% score intervals do not overlap. That is the clearest overall quality signal in the current public data.
Which is better for coding, GPT-5.4 or Trinity-Large-Thinking?
GPT-5.4 leads the public coding lane, 49.7 to 21, with Supported evidence for both models and non-overlapping 90% intervals.
Which is better for agentic tasks, GPT-5.4 or Trinity-Large-Thinking?
GPT-5.4 scores higher for agentic tasks on the public lane, 47.8 to 17.7. Trinity-Large-Thinking is scored on Estimated evidence for agentic tasks, so the reading is directional rather than like-for-like. BenchLM does not name a winner for agentic tasks from a directional reading; read the shared benchmark rows directly and test the models on the same task set.
Which costs less, GPT-5.4 or Trinity-Large-Thinking?
For the stated presets, chat costs $0.01 on GPT-5.4 and $0.0007 on Trinity-Large-Thinking; repository review costs $0.17 and $0.0152; the cache-heavy agent loop costs $0.25 and $0.064. Trinity-Large-Thinking has no published cached-input rate, so cached tokens use its listed input rate.
Which has the larger context window, GPT-5.4 or Trinity-Large-Thinking?
GPT-5.4 has the larger documented context window: 1.05M, compared with 512K.
Benchmark evidence
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Browse raw public benchmark evidence41 rows
Agentic
Terminal-Bench 2.0
GPT-5.475.1%SourceTrinity-Large-Thinking—Not directly comparable
CyberGym
GPT-5.479.0%SourceTrinity-Large-Thinking—Not directly comparable
BrowseComp
GPT-5.482.7%SourceTrinity-Large-Thinking—Not directly comparable
OSWorld-Verified
GPT-5.475%SourceTrinity-Large-Thinking—Not directly comparable
MCP Atlas
GPT-5.470.6%SourceTrinity-Large-Thinking—Not directly comparable
Toolathlon
GPT-5.454.6%SourceTrinity-Large-Thinking—Not directly comparable
τ²-bench results
GPT-5.498.9%SourceTrinity-Large-Thinking—Not directly comparable
Claw-Eval
GPT-5.460.3%SourceTrinity-Large-Thinking—Not directly comparable
DeepSearchQA
GPT-5.473.6%SourceTrinity-Large-Thinking—Not directly comparable
Gert Labs
Shared sourceGPT-5.464.89%Trinity-Large-Thinking32.55%GPT-5.4 leads this result
ResearchClawBench
GPT-5.415.3%SourceTrinity-Large-Thinking—Not directly comparable
JobBench
GPT-5.438.9%SourceTrinity-Large-Thinking—Not directly comparable
ExploitGym
GPT-5.46.0%SourceTrinity-Large-Thinking—Not directly comparable
ApprenticeBench
GPT-5.411%SourceTrinity-Large-Thinking—Not directly comparable
Coding
LiveCodeBench Pro
GPT-5.487.5%SourceTrinity-Large-Thinking—Not directly comparable
SWE-bench Pro
GPT-5.457.7%SourceTrinity-Large-Thinking—Not directly comparable
React Native Evals
GPT-5.485.3%SourceTrinity-Large-Thinking—Not directly comparable
Vibe Code Bench
GPT-5.467.42%SourceTrinity-Large-Thinking—Not directly comparable
SWE-bench Verified*
GPT-5.4—Trinity-Large-Thinking63.2%SourceNot directly comparable
Reasoning
Multimodal
MMMU-Pro
GPT-5.481.2%SourceTrinity-Large-Thinking—Not directly comparable
OfficeQA Pro
GPT-5.453.2%SourceTrinity-Large-Thinking—Not directly comparable
MMMU-Pro w/ Python
GPT-5.482.1%SourceTrinity-Large-Thinking—Not directly comparable
CharXiv
GPT-5.482.8%SourceTrinity-Large-Thinking—Not directly comparable
ERQA
GPT-5.465.4%SourceTrinity-Large-Thinking—Not directly comparable
SimpleVQA
GPT-5.461.1%SourceTrinity-Large-Thinking—Not directly comparable
ScreenSpot Pro
GPT-5.485.4%SourceTrinity-Large-Thinking—Not directly comparable
ZeroBench
GPT-5.441.0%SourceTrinity-Large-Thinking—Not directly comparable
MedXpertQA (MM)
GPT-5.477.1%SourceTrinity-Large-Thinking—Not directly comparable
Knowledge
GPQA
GPT-5.492.8%SourceTrinity-Large-Thinking—Not directly comparable
HLE
GPT-5.452.1%SourceTrinity-Large-Thinking—Not directly comparable
HLE w/o tools
GPT-5.439.8%SourceTrinity-Large-Thinking—Not directly comparable
GPQA-D
GPT-5.492.8%SourceTrinity-Large-Thinking76.3%SourceGPT-5.4 leads this result
HealthBench Hard
GPT-5.440.1%SourceTrinity-Large-Thinking—Not directly comparable
MedXpertQA (Text)
GPT-5.459.6%SourceTrinity-Large-Thinking—Not directly comparable
HealthBench Professional
GPT-5.448.1%SourceTrinity-Large-Thinking—Not directly comparable
MMLU-Pro (Arcee)
GPT-5.4—Trinity-Large-Thinking83.4%SourceNot directly comparable
Math
41 public results · 2 shared
Watch GPT-5.4 vs Trinity-Large-Thinking
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.
Last updated September 29, 2026