Math
Like-for-like- GPT-5.4 mini
- 21.7
- GPT-5.6 Terra
- 80.8
- Weighted basis
- 2 vs 2 rows
- Reading
- GPT-5.6 Terra leads
Model comparison
GPT-5.4 mini vs GPT-5.6 Terra
Updated August 10, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
GPT-5.6 Terra has the higher public score estimate, 72.29 versus 55.93, but the 90% score intervals overlap. Treat that as a lead, not a settled winner.
7 results are shared. Category rows based on different benchmark sets are marked directional and do not name a winner.
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-5.6 Terra
GPT-5.6 Terra has the larger documented context window.
Confidence: documented
Chat turn cost
1K fresh input + 500 output tokens
GPT-5.4 mini
GPT-5.4 mini 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
GPT-5.4 mini
GPT-5.4 mini 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
Repository review cost
50K fresh input + 3K output tokens
GPT-5.4 mini
GPT-5.4 mini has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Coding work
Code generation, repair, and software-engineering tasks
Not enough matched evidence
No shared weighted benchmark basis supports a winner.
Confidence: limited
Agentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
The category averages use different weighted benchmark sets, so they are directional rather than like-for-like.
Confidence: limited
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
7 GPT-5.4 mini only
7 Shared
15 GPT-5.6 Terra only
- Shared results
- 7
- GPT-5.4 mini only
- 7
- GPT-5.6 Terra only
- 15
- Like-for-like categories
- 2 / 8
2 categories use different evidence sets. Those rows remain visible for coverage context but do not name a winner.
Category results, on a stated basis
Each row states whether both averages use the same weighted benchmark set. Directional and not-comparable rows remain visible, but they never receive a winner in this template.
Multimodal
Like-for-like- GPT-5.4 mini
- 76.6
- GPT-5.6 Terra
- 80.7
- Weighted basis
- 1 vs 1 rows
- Reading
- GPT-5.6 Terra leads
Agentic
Directional only- GPT-5.4 mini
- 65.7
- GPT-5.6 Terra
- 87.4
- Weighted basis
- 2 vs 2 rows
- Reading
- Directional only
Knowledge
Directional only- GPT-5.4 mini
- 47.8
- GPT-5.6 Terra
- 92.9
- Weighted basis
- 2 vs 1 rows
- Reading
- Directional only
Coding
Not comparable- GPT-5.4 mini
- Not measured
- GPT-5.6 Terra
- 63.4
- Weighted basis
- 0 vs 1 rows
- Reading
- Not comparable
Reasoning
Not comparable- GPT-5.4 mini
- Not measured
- GPT-5.6 Terra
- 83.9
- Weighted basis
- 0 vs 1 rows
- Reading
- Not comparable
Multilingual
Not comparable- GPT-5.4 mini
- Not measured
- GPT-5.6 Terra
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Instruction following
Not comparable- GPT-5.4 mini
- Not measured
- GPT-5.6 Terra
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
| Category | GPT-5.4 mini | GPT-5.6 Terra | Weighted basis | Reading |
|---|---|---|---|---|
| Math | 21.7 | 80.8 | Like-for-like2 vs 2 rows | GPT-5.6 Terra leads |
| Multimodal | 76.6 | 80.7 | Like-for-like1 vs 1 rows | GPT-5.6 Terra leads |
| Agentic | 65.7 | 87.4 | Directional only2 vs 2 rows | Directional only |
| Knowledge | 47.8 | 92.9 | Directional only2 vs 1 rows | Directional only |
| Coding | Not measured | 63.4 | Not comparable0 vs 1 rows | Not comparable |
| Reasoning | Not measured | 83.9 | Not comparable0 vs 1 rows | Not comparable |
| Multilingual | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Instruction following | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
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.
FrontierMath v2 (Tier 4)
Math
GPT-5.4 mini: 2.080%GPT-5.6 Terra: 68.300%Normalized gap 66.2GPT-5.4 mini source GPT-5.6 Terra sourceFrontierMath v2 (Tiers 1-3)
Math
GPT-5.4 mini: 28.280%GPT-5.6 Terra: 84.900%Normalized gap 56.6GPT-5.4 mini source GPT-5.6 Terra sourceTerminal-Bench 2.0
Agentic
GPQA
Knowledge
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-5.4 mini
- $0.003
- Fits in one request
- GPT-5.6 Terra
- $0.008
- Fits in one request
GPT-5.4 mini has the lower modeled cost
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- GPT-5.4 mini
- $0.051
- Fits in one request
- GPT-5.6 Terra
- $0.136
- Fits in one request
GPT-5.4 mini 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 mini
- $0.075
- Fits in one request
- GPT-5.6 Terra
- $0.2
- Fits in one request
GPT-5.4 mini has the lower modeled cost
Costs use the listed standard API rates.
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationGPT-5.4 miniGPT-5.6 Terra
Context window
Maximum documented context; output-token limits may be lower.
GPT-5.4 mini
GPT-5.6 Terra
1.05M
OpenAI model catalogAPI model ID
GPT-5.4 mini
gpt-5.4-mini
OpenAI GPT-5.4 mini model documentationGPT-5.6 Terra
gpt-5.6-terra
OpenAI model catalogCached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
GPT-5.4 mini
$0.075 per 1M cached input tokens
OpenAI pricingGPT-5.6 Terra
$0.2 per 1M cached input tokens
OpenAI pricingDocumented inputs
GPT-5.4 mini
text, image
OpenAI model catalogGPT-5.6 Terra
text, image
OpenAI model catalogDocumented outputs
GPT-5.4 mini
GPT-5.6 Terra
Provider availability
GPT-5.4 mini
Generally Available · OpenAI Responses API
OpenAI model catalogGPT-5.6 Terra
Generally Available · OpenAI Responses API
OpenAI model catalogReasoning profile
GPT-5.4 mini
Reasoning
GPT-5.6 Terra
Reasoning
Weight access
GPT-5.4 mini
Proprietary
GPT-5.6 Terra
Proprietary
License
GPT-5.4 mini
Proprietary
GPT-5.6 Terra
Proprietary
Release date
GPT-5.4 mini
2026-03-17
GPT-5.6 Terra
2026-07-09
If you already use one of these models
- Deployment change
- Both entries list OpenAI as the provider. Confirm endpoint, model ID, limits, and feature support before switching.
- Quality signal
- GPT-5.6 Terra has the higher public score estimate, 72.29 versus 55.93, but the 90% score intervals overlap.
- Workload cost
- Repository review: $0.051 vs $0.136. Cache-heavy agent loop: $0.075 vs $0.2.
- Context tradeoff
- GPT-5.6 Terra has the larger documented window (1.05M).
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 evidence29 rows
Agentic
Terminal-Bench 2.0
GPT-5.4 mini60%SourceGPT-5.6 Terra87.4%SourceGPT-5.6 Terra leads this result
OSWorld-Verified
GPT-5.4 mini72.1%SourceGPT-5.6 Terra—Not directly comparable
MCP Atlas
GPT-5.4 mini57.7%SourceGPT-5.6 Terra—Not directly comparable
Toolathlon
GPT-5.4 mini42.9%SourceGPT-5.6 Terra53.1%SourceGPT-5.6 Terra leads this result
τ²-bench results
GPT-5.4 mini93.4%SourceGPT-5.6 Terra—Not directly comparable
BrowseComp
GPT-5.4 mini—GPT-5.6 Terra87.5%SourceNot directly comparable
OSWorld 2.0
GPT-5.4 mini—GPT-5.6 Terra50.2%SourceNot directly comparable
CyberGym
GPT-5.4 mini—GPT-5.6 Terra81.8%SourceNot directly comparable
ExploitGym
GPT-5.4 mini—GPT-5.6 Terra23.2%SourceNot directly comparable
Coding
Vibe Code Bench
GPT-5.4 mini47.97%SourceGPT-5.6 Terra—Not directly comparable
FrontierCode 1.1 Main
GPT-5.4 mini27.0%SourceGPT-5.6 Terra—Not directly comparable
SWE-bench Pro
GPT-5.4 mini—GPT-5.6 Terra63.4%SourceNot directly comparable
Terminal-Bench 2.0
GPT-5.4 mini—GPT-5.6 Terra87.4%SourceNot directly comparable
deepSwe
GPT-5.4 mini—GPT-5.6 Terra69.6%SourceNot directly comparable
FrontierCode 1.1 Extended
GPT-5.4 mini—GPT-5.6 Terra55.8%SourceNot directly comparable
cursorBench32
GPT-5.4 mini—GPT-5.6 Terra64.9%SourceNot directly comparable
Reasoning
Knowledge
GPQA
GPT-5.4 mini88%SourceGPT-5.6 Terra92.9%SourceGPT-5.6 Terra leads this result
HLE
GPT-5.4 mini41.5%SourceGPT-5.6 Terra—Not directly comparable
HLE w/o tools
GPT-5.4 mini28.2%SourceGPT-5.6 Terra—Not directly comparable
GPQA-D
GPT-5.4 mini—GPT-5.6 Terra92.9%SourceNot directly comparable
HealthBench Professional
GPT-5.4 mini—GPT-5.6 Terra57.7%SourceNot directly comparable
HealthBench Hard
GPT-5.4 mini—GPT-5.6 Terra32.7%SourceNot directly comparable
Math
FrontierMath v2 (Tiers 1-3)
GPT-5.4 mini28.280%SourceGPT-5.6 Terra84.900%SourceGPT-5.6 Terra leads this result
FrontierMath v2 (Tier 4)
GPT-5.4 mini2.080%SourceGPT-5.6 Terra68.300%SourceGPT-5.6 Terra leads this result
FrontierMath (legacy)
GPT-5.4 mini—GPT-5.6 Terra84.9%SourceNot directly comparable
Multimodal
Frequently asked questions
Which is better, GPT-5.4 mini or GPT-5.6 Terra?
GPT-5.6 Terra has the higher public score estimate, 72.29 versus 55.93, 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-5.4 mini or GPT-5.6 Terra?
The published evidence does not provide a shared weighted coding basis for both models, so BenchLM does not name a coding winner.
Which is better for agentic tasks, GPT-5.4 mini or GPT-5.6 Terra?
The current agentic tasks averages use different weighted benchmark sets, so BenchLM does not name a winner from them. Read the shared benchmark rows directly and test the models on the same task set.
Which costs less, GPT-5.4 mini or GPT-5.6 Terra?
For the stated presets, chat costs $0.003 on GPT-5.4 mini and $0.008 on GPT-5.6 Terra; repository review costs $0.051 and $0.136; the cache-heavy agent loop costs $0.075 and $0.2. Costs use the listed standard API rates.
Which has the larger context window, GPT-5.4 mini or GPT-5.6 Terra?
GPT-5.6 Terra has the larger documented context window: 1.05M, compared with 400K.
Related comparisons
Last updated August 10, 2026
Watch GPT-5.4 mini vs GPT-5.6 Terra
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.