Coding
Like-for-like- Gemini 3.5 Flash
- 55.1
- GPT-5.4
- 57.7
- Weighted basis
- 1 vs 1 rows
- Reading
- GPT-5.4 leads
Five or fewer confirmed AI changes, with original sources, on mornings when something changed.A free source-linked morning brief.
Start free briefUpdated August 18, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
Decision reading
GPT-5.4 has the higher public score estimate, 73.36 versus 64.67, but the 90% score intervals overlap. Treat that as a lead, not a settled winner.
16 results are shared. Category rows based on different benchmark sets are marked directional and do not name a winner.
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.
Code generation, repair, and software-engineering tasks
GPT-5.4
GPT-5.4 leads on the same 1 weighted benchmark row.
Confidence: limited
Prompts that approach the documented context limit
GPT-5.4
GPT-5.4 has the larger documented context window.
Confidence: documented
1K fresh input + 500 output tokens
Gemini 3.5 Flash
Gemini 3.5 Flash has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
200K cached + 20K fresh input + 10K output tokens
Gemini 3.5 Flash
Gemini 3.5 Flash has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
50K fresh input + 3K output tokens
Gemini 3.5 Flash
Gemini 3.5 Flash has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
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
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
4 categories use different evidence sets. Those rows remain visible for coverage context but do not name a winner.
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.
| Category | Gemini 3.5 Flash | GPT-5.4 | Weighted basis | Reading |
|---|---|---|---|---|
| Coding | 55.1 | 57.7 | Like-for-like1 vs 1 rows | GPT-5.4 leads |
| Math | 32.9 | 42.5 | Like-for-like2 vs 2 rows | GPT-5.4 leads |
| Agentic | 77.2 | 77.2 | Directional only2 vs 3 rows | Directional only |
| Reasoning | 74.7 | 74.0 | Directional only2 vs 1 rows | Directional only |
| Knowledge | 40.2 | 57.6 | Directional only1 vs 2 rows | Directional only |
| Multimodal | 83.8 | 73.2 | Directional only2 vs 3 rows | Directional only |
| 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 |
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
HLE
Knowledge
FrontierMath v2 (Tiers 1-3)
Math
OSWorld-Verified
Agentic
SWE-bench Pro
Coding
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.
1K fresh input + 500 output tokens
Gemini 3.5 Flash has the lower modeled cost
Costs use the listed standard API rates.
50K fresh input + 3K output tokens
Gemini 3.5 Flash has the lower modeled cost
Costs use the listed standard API rates.
200K cached + 20K fresh input + 10K output tokens
Gemini 3.5 Flash has the lower modeled cost
Costs use the listed standard API rates.
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
Maximum documented context; output-token limits may be lower.
Gemini 3.5 Flash
GPT-5.4
1.05M
OpenAI pricingGemini 3.5 Flash
gemini-3.5-flash
Google Gemini API pricingGPT-5.4
gpt-5.4
OpenAI pricingA missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Gemini 3.5 Flash
$0.15 per 1M cached input tokens
Google Gemini API pricingGPT-5.4
$0.25 per 1M cached input tokens
OpenAI pricingGemini 3.5 Flash
Not sourced
GPT-5.4
Not sourced
Gemini 3.5 Flash
Not sourced
GPT-5.4
Not sourced
Gemini 3.5 Flash
Not sourced
GPT-5.4
Not sourced
Gemini 3.5 Flash
Reasoning
GPT-5.4
Reasoning
Gemini 3.5 Flash
Proprietary
GPT-5.4
Proprietary
Gemini 3.5 Flash
Proprietary
GPT-5.4
Proprietary
Gemini 3.5 Flash
2026-05-19
GPT-5.4
2026-03-05
Run the same representative tasks against both endpoints before changing production traffic.
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Terminal-Bench 2.0
Gemini 3.5 Flash leads this result
MCP Atlas
Gemini 3.5 Flash leads this result
Toolathlon
Gemini 3.5 Flash leads this result
OSWorld-Verified
Gemini 3.5 Flash leads this result
Finance Agent v2
Not directly comparable
Gert Labs
Shared sourceGPT-5.4 leads this result
ResearchClawBench
Shared sourceGemini 3.5 Flash leads this result
CyberGym
Not directly comparable
BrowseComp
Not directly comparable
τ²-bench results
Not directly comparable
Claw-Eval
Not directly comparable
DeepSearchQA
Not directly comparable
JobBench
Not directly comparable
ExploitGym
Not directly comparable
Terminal-Bench 2.0
Not directly comparable
SWE-bench Pro
GPT-5.4 leads this result
Vibe Code Bench
Shared sourceGPT-5.4 leads this result
cursorBench31
Not directly comparable
cursorBench32
Not directly comparable
EEBench
Shared sourceGPT-5.4 leads this result
LiveCodeBench Pro
Not directly comparable
React Native Evals
Not directly comparable
GPQA-D
GPT-5.4 leads this result
HLE
GPT-5.4 leads this result
GPQA
Not directly comparable
HLE w/o tools
Not directly comparable
HealthBench Hard
Not directly comparable
MedXpertQA (Text)
Not directly comparable
HealthBench Professional
Not directly comparable
FrontierMath v2 (Tiers 1-3)
Shared sourceGPT-5.4 leads this result
FrontierMath v2 (Tier 4)
Shared sourceGPT-5.4 leads this result
CharXiv
Gemini 3.5 Flash leads this result
MMMU-Pro
Gemini 3.5 Flash leads this result
Blueprint-Bench 2
Not directly comparable
OfficeQA Pro
Not directly comparable
MMMU-Pro w/ Python
Not directly comparable
ERQA
Not directly comparable
SimpleVQA
Not directly comparable
ScreenSpot Pro
Not directly comparable
ZeroBench
Not directly comparable
MedXpertQA (MM)
Not directly comparable
GPT-5.4 has the higher public score estimate, 73.36 versus 64.67, but the 90% score intervals overlap. The higher estimate is not a decisive winner because the uncertainty ranges overlap.
GPT-5.4 leads the like-for-like coding comparison across 1 shared weighted benchmark row.
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.
For the stated presets, chat costs $0.006 on Gemini 3.5 Flash and $0.01 on GPT-5.4; repository review costs $0.102 and $0.17; the cache-heavy agent loop costs $0.15 and $0.25. Costs use the listed standard API rates.
GPT-5.4 has the larger documented context window: 1.05M, compared with 1M.
Last updated August 18, 2026
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.