Agentic
Not comparable- GPT-5.2-Codex
- Not measured
- Qwen3.5 397B
- 56.5
- Weighted basis
- 0 vs 2 rows
- Reading
- Not comparable
GPT-5.2-Codex vs Qwen3.5 397B
Updated September 2, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
Decision reading
GPT-5.2-Codex and Qwen3.5 397B have the same public score on the current data.
1 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-5.2-Codex
GPT-5.2-Codex has the larger documented context window.
Confidence: documented
Chat turn cost
1K fresh input + 500 output tokens
Qwen3.5 397B
Qwen3.5 397B 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
Qwen3.5 397B
Qwen3.5 397B 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
No shared weighted benchmark basis supports a winner.
Confidence: limited
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Not enough matched evidence
The page does not recommend a cost winner because at least one model cannot fit the stated workload in one request. Qwen3.5 397B does not fit this workload in one request. GPT-5.2-Codex has no published cached-input rate, so cached tokens use its listed input rate. Qwen3.5 397B has no published cached-input rate, so cached tokens use its listed input rate.
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.
2 GPT-5.2-Codex only
1 Shared
37 Qwen3.5 397B only
- Shared results
- 1
- GPT-5.2-Codex only
- 2
- Qwen3.5 397B only
- 37
- Like-for-like categories
- 0 / 8
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.
Coding
Not comparable- GPT-5.2-Codex
- Not measured
- Qwen3.5 397B
- 66.5
- Weighted basis
- 0 vs 2 rows
- Reading
- Not comparable
Reasoning
Not comparable- GPT-5.2-Codex
- Not measured
- Qwen3.5 397B
- 63.2
- Weighted basis
- 0 vs 1 rows
- Reading
- Not comparable
Knowledge
Not comparable- GPT-5.2-Codex
- Not measured
- Qwen3.5 397B
- 56.6
- Weighted basis
- 0 vs 4 rows
- Reading
- Not comparable
Math
Not comparable- GPT-5.2-Codex
- Not measured
- Qwen3.5 397B
- 90.6
- Weighted basis
- 0 vs 2 rows
- Reading
- Not comparable
Multilingual
Not comparable- GPT-5.2-Codex
- Not measured
- Qwen3.5 397B
- 84.7
- Weighted basis
- 0 vs 1 rows
- Reading
- Not comparable
Multimodal
Not comparable- GPT-5.2-Codex
- Not measured
- Qwen3.5 397B
- 79.6
- Weighted basis
- 0 vs 2 rows
- Reading
- Not comparable
Instruction following
Not comparable- GPT-5.2-Codex
- Not measured
- Qwen3.5 397B
- 92.6
- Weighted basis
- 0 vs 1 rows
- Reading
- Not comparable
| Category | GPT-5.2-Codex | Qwen3.5 397B | Weighted basis | Reading |
|---|---|---|---|---|
| Agentic | Not measured | 56.5 | Not comparable0 vs 2 rows | Not comparable |
| Coding | Not measured | 66.5 | Not comparable0 vs 2 rows | Not comparable |
| Reasoning | Not measured | 63.2 | Not comparable0 vs 1 rows | Not comparable |
| Knowledge | Not measured | 56.6 | Not comparable0 vs 4 rows | Not comparable |
| Math | Not measured | 90.6 | Not comparable0 vs 2 rows | Not comparable |
| Multilingual | Not measured | 84.7 | Not comparable0 vs 1 rows | Not comparable |
| Multimodal | Not measured | 79.6 | Not comparable0 vs 2 rows | Not comparable |
| Instruction following | Not measured | 92.6 | Not comparable0 vs 1 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.
A shared-evidence shape is not available.
BenchLM does not draw a radar or infer missing axes when the matched evidence is too sparse.
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-5.2-Codex
- $0.00875
- Fits in one request
- Qwen3.5 397B
- $0.0024
- Fits in one request
Qwen3.5 397B has the lower modeled cost
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- GPT-5.2-Codex
- $0.1295
- Fits in one request
- Qwen3.5 397B
- $0.0408
- Fits in one request
Qwen3.5 397B 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-5.2-Codex
- $0.525
- Fits in one request
- Cached input priced at the published list-input rate
- Qwen3.5 397B
- $0.168
- Does not fit in one request
- Cached input priced at the published list-input rate
Qwen3.5 397B does not fit this workload in one request. GPT-5.2-Codex has no published cached-input rate, so cached tokens use its listed input rate. Qwen3.5 397B 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-5.2-CodexQwen3.5 397B
Context window
Maximum documented context; output-token limits may be lower.
GPT-5.2-Codex
400K
Qwen3.5 397B
128K
API model ID
GPT-5.2-Codex
Not sourced
Qwen3.5 397B
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
GPT-5.2-Codex
Not published
Qwen3.5 397B
Not published
Documented inputs
GPT-5.2-Codex
Not sourced
Qwen3.5 397B
Not sourced
Documented outputs
GPT-5.2-Codex
Not sourced
Qwen3.5 397B
Not sourced
Provider availability
GPT-5.2-Codex
Not sourced
Qwen3.5 397B
Not sourced
Reasoning profile
GPT-5.2-Codex
Reasoning
Qwen3.5 397B
Non-Reasoning
Weight access
GPT-5.2-Codex
Proprietary
Qwen3.5 397B
Open Weight
License
GPT-5.2-Codex
Proprietary
Qwen3.5 397B
Open Weight
Release date
GPT-5.2-Codex
2025-12-18
Qwen3.5 397B
2026-02-16
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-5.2-Codex and Qwen3.5 397B have the same public score on the current data.
- Workload cost
- Repository review: $0.1295 vs $0.0408. Cache-heavy agent loop: $0.525 vs $0.168.
- Context tradeoff
- GPT-5.2-Codex has the larger documented window (400K).
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 evidence40 rows
Agentic
Gert Labs
Shared sourceGPT-5.2-Codex51.79%Qwen3.5 397B46.76%GPT-5.2-Codex leads this result
JobBench
GPT-5.2-Codex26.0%SourceQwen3.5 397B—Not directly comparable
Terminal-Bench 2.0
GPT-5.2-Codex—Qwen3.5 397B52.5%SourceNot directly comparable
BrowseComp
GPT-5.2-Codex—Qwen3.5 397B62%SourceNot directly comparable
Claw-Eval
GPT-5.2-Codex—Qwen3.5 397B56.8%SourceNot directly comparable
QwenClawBench
GPT-5.2-Codex—Qwen3.5 397B51.8%SourceNot directly comparable
τ³-bench results
GPT-5.2-Codex—Qwen3.5 397B68.4%SourceNot directly comparable
VITA-Bench
GPT-5.2-Codex—Qwen3.5 397B43.7%SourceNot directly comparable
DeepPlanning
GPT-5.2-Codex—Qwen3.5 397B37.6%SourceNot directly comparable
Toolathlon
GPT-5.2-Codex—Qwen3.5 397B36.3%SourceNot directly comparable
MCP Atlas
GPT-5.2-Codex—Qwen3.5 397B46.1%SourceNot directly comparable
MCP-Tasks
GPT-5.2-Codex—Qwen3.5 397B74.2%SourceNot directly comparable
WideResearch
GPT-5.2-Codex—Qwen3.5 397B74.0%SourceNot directly comparable
ResearchClawBench
GPT-5.2-Codex—Qwen3.5 397B14.2%SourceNot directly comparable
Coding
Vibe Code Bench
GPT-5.2-Codex37.91%SourceQwen3.5 397B—Not directly comparable
SWE-bench Verified
GPT-5.2-Codex—Qwen3.5 397B76.2%SourceNot directly comparable
LiveCodeBench v6
GPT-5.2-Codex—Qwen3.5 397B83.6%SourceNot directly comparable
SWE-bench Pro
GPT-5.2-Codex—Qwen3.5 397B50.9%SourceNot directly comparable
Reasoning
Knowledge
GPQA
GPT-5.2-Codex—Qwen3.5 397B88.4%SourceNot directly comparable
SuperGPQA
GPT-5.2-Codex—Qwen3.5 397B70.4%SourceNot directly comparable
MMLU-Pro
GPT-5.2-Codex—Qwen3.5 397B87.8%SourceNot directly comparable
MMLU-Redux
GPT-5.2-Codex—Qwen3.5 397B94.9%SourceNot directly comparable
C-Eval
GPT-5.2-Codex—Qwen3.5 397B93%SourceNot directly comparable
HLE
GPT-5.2-Codex—Qwen3.5 397B28.7%SourceNot directly comparable
Math
AIME26
GPT-5.2-Codex—Qwen3.5 397B93.3%SourceNot directly comparable
HMMT Feb 2025
GPT-5.2-Codex—Qwen3.5 397B94.8%SourceNot directly comparable
HMMT Nov 2025
GPT-5.2-Codex—Qwen3.5 397B92.7%SourceNot directly comparable
HMMT Feb 2026
GPT-5.2-Codex—Qwen3.5 397B87.9%SourceNot directly comparable
MMAnswerBench
GPT-5.2-Codex—Qwen3.5 397B80.9%SourceNot directly comparable
Multilingual
Multimodal
MMMU-Pro
GPT-5.2-Codex—Qwen3.5 397B79%SourceNot directly comparable
MathVision
GPT-5.2-Codex—Qwen3.5 397B88.6%SourceNot directly comparable
CharXiv
GPT-5.2-Codex—Qwen3.5 397B80.8%SourceNot directly comparable
VideoMMMU
GPT-5.2-Codex—Qwen3.5 397B84.7%SourceNot directly comparable
ScreenSpot Pro
GPT-5.2-Codex—Qwen3.5 397B65.6%SourceNot directly comparable
V*
GPT-5.2-Codex—Qwen3.5 397B95.8%SourceNot directly comparable
Instruction following
IFEval
GPT-5.2-Codex—Qwen3.5 397B92.6%SourceNot directly comparable
Frequently asked questions
Which is better, GPT-5.2-Codex or Qwen3.5 397B?
GPT-5.2-Codex and Qwen3.5 397B have the same public score on the current data. The page therefore keeps the decision tied to the specific documented workload.
Which is better for coding, GPT-5.2-Codex or Qwen3.5 397B?
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-5.2-Codex or Qwen3.5 397B?
The published evidence does not provide a shared weighted agentic tasks basis for both models, so BenchLM does not name a agentic tasks winner.
Which costs less, GPT-5.2-Codex or Qwen3.5 397B?
For the stated presets, chat costs $0.00875 on GPT-5.2-Codex and $0.0024 on Qwen3.5 397B; repository review costs $0.1295 and $0.0408; the cache-heavy agent loop costs $0.525 and $0.168. Qwen3.5 397B does not fit this workload in one request. GPT-5.2-Codex has no published cached-input rate, so cached tokens use its listed input rate. Qwen3.5 397B has no published cached-input rate, so cached tokens use its listed input rate.
Which has the larger context window, GPT-5.2-Codex or Qwen3.5 397B?
GPT-5.2-Codex has the larger documented context window: 400K, compared with 128K.
Related comparisons
Last updated September 2, 2026
Watch GPT-5.2-Codex vs Qwen3.5 397B
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.