Skip to main content
Radar

Provider changes are easy to miss. Radar watches releases, pricing, deprecations, and incidents at the source.Provider changes are easy to miss.

See Radar
BenchLM
Advertise

Model comparison

GPT-5.6 Terra vs Qwen3.5 397B

Updated August 10, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.

GPT-5.6 Terra

OpenAI

72.3/100

Estimated · Public rank #12

90% interval 62.6–82.0

Qwen3.5 397B

Alibaba

56.2/100

Estimated · Public rank #81

90% interval 44.7–67.8

GPT-5.6 Terra has the higher public score estimate, 72.29 versus 56.24, but the 90% score intervals overlap. Treat that as a lead, not a settled winner.

6 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.

  • Agentic work

    Tool use, computer use, and multi-step task completion

    GPT-5.6 Terra

    GPT-5.6 Terra leads on the same 2 weighted benchmark rows.

    Confidence: limited

  • 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

    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

    The category averages use different weighted benchmark sets, so they are directional rather than like-for-like.

    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. 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.

Shared results
6
GPT-5.6 Terra only
16
Qwen3.5 397B only
32
Like-for-like categories
1 / 8

3 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.

Agentic

Like-for-like
GPT-5.6 Terra
87.4
Qwen3.5 397B
56.5
Weighted basis
2 vs 2 rows
Reading
GPT-5.6 Terra leads

Coding

Directional only
GPT-5.6 Terra
63.4
Qwen3.5 397B
66.5
Weighted basis
1 vs 2 rows
Reading
Directional only

Knowledge

Directional only
GPT-5.6 Terra
92.9
Qwen3.5 397B
56.6
Weighted basis
1 vs 4 rows
Reading
Directional only

Multimodal

Directional only
GPT-5.6 Terra
80.7
Qwen3.5 397B
79.6
Weighted basis
1 vs 2 rows
Reading
Directional only

Reasoning

Not comparable
GPT-5.6 Terra
83.9
Qwen3.5 397B
63.2
Weighted basis
1 vs 1 rows
Reading
Not comparable

Math

Not comparable
GPT-5.6 Terra
80.8
Qwen3.5 397B
90.6
Weighted basis
2 vs 2 rows
Reading
Not comparable

Multilingual

Not comparable
GPT-5.6 Terra
Not measured
Qwen3.5 397B
84.7
Weighted basis
0 vs 1 rows
Reading
Not comparable

Instruction following

Not comparable
GPT-5.6 Terra
Not measured
Qwen3.5 397B
92.6
Weighted basis
0 vs 1 rows
Reading
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.

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-5.6 Terra
$0.008
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-5.6 Terra
$0.136
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-5.6 Terra
$0.2
Fits in one request
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. 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.

Context window

Maximum documented context; output-token limits may be lower.

GPT-5.6 Terra

1.05M

OpenAI model catalog

Qwen3.5 397B

128K

Cached-input rate

A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.

GPT-5.6 Terra

$0.2 per 1M cached input tokens

OpenAI pricing

Qwen3.5 397B

Not published

Provider availability

GPT-5.6 Terra

Generally Available · OpenAI Responses API

OpenAI model catalog

Qwen3.5 397B

Not sourced

Reasoning profile

GPT-5.6 Terra

Reasoning

Qwen3.5 397B

Non-Reasoning

Weight access

GPT-5.6 Terra

Proprietary

Qwen3.5 397B

Open Weight

License

GPT-5.6 Terra

Proprietary

Qwen3.5 397B

Open Weight

Release date

GPT-5.6 Terra

2026-07-09

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-5.6 Terra has the higher public score estimate, 72.29 versus 56.24, but the 90% score intervals overlap.
Workload cost
Repository review: $0.136 vs $0.0408. Cache-heavy agent loop: $0.2 vs $0.168.
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 evidence54 rows

Agentic

  • Terminal-Bench 2.0

    GPT-5.6 Terra87.4%
    Source
    Qwen3.5 397B52.5%
    Source

    GPT-5.6 Terra leads this result

  • BrowseComp

    GPT-5.6 Terra87.5%
    Source
    Qwen3.5 397B62%
    Source

    GPT-5.6 Terra leads this result

  • OSWorld 2.0

    GPT-5.6 Terra50.2%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • CyberGym

    GPT-5.6 Terra81.8%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • ExploitGym

    GPT-5.6 Terra23.2%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • Toolathlon

    GPT-5.6 Terra53.1%
    Source
    Qwen3.5 397B36.3%
    Source

    GPT-5.6 Terra leads this result

  • Claw-Eval

    GPT-5.6 Terra—
    Qwen3.5 397B56.8%
    Source

    Not directly comparable

  • QwenClawBench

    GPT-5.6 Terra—
    Qwen3.5 397B51.8%
    Source

    Not directly comparable

  • τ³-bench results

    GPT-5.6 Terra—
    Qwen3.5 397B68.4%
    Source

    Not directly comparable

  • VITA-Bench

    GPT-5.6 Terra—
    Qwen3.5 397B43.7%
    Source

    Not directly comparable

  • DeepPlanning

    GPT-5.6 Terra—
    Qwen3.5 397B37.6%
    Source

    Not directly comparable

  • MCP Atlas

    GPT-5.6 Terra—
    Qwen3.5 397B46.1%
    Source

    Not directly comparable

  • MCP-Tasks

    GPT-5.6 Terra—
    Qwen3.5 397B74.2%
    Source

    Not directly comparable

  • WideResearch

    GPT-5.6 Terra—
    Qwen3.5 397B74.0%
    Source

    Not directly comparable

  • Gert Labs

    GPT-5.6 Terra—
    Qwen3.5 397B46.76%
    Source

    Not directly comparable

  • ResearchClawBench

    GPT-5.6 Terra—
    Qwen3.5 397B14.2%
    Source

    Not directly comparable

Coding

  • SWE-bench Pro

    GPT-5.6 Terra63.4%
    Source
    Qwen3.5 397B50.9%
    Source

    GPT-5.6 Terra leads this result

  • Terminal-Bench 2.0

    GPT-5.6 Terra87.4%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • deepSwe

    GPT-5.6 Terra69.6%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • FrontierCode 1.1 Extended

    GPT-5.6 Terra55.8%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • cursorBench32

    GPT-5.6 Terra64.9%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • SWE-bench Verified

    GPT-5.6 Terra—
    Qwen3.5 397B76.2%
    Source

    Not directly comparable

  • LiveCodeBench v6

    GPT-5.6 Terra—
    Qwen3.5 397B83.6%
    Source

    Not directly comparable

Reasoning

  • ARC-AGI-2

    GPT-5.6 Terra83.9%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • ARC-AGI-3

    GPT-5.6 Terra0.8%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • LongBench v2

    GPT-5.6 Terra—
    Qwen3.5 397B63.2%
    Source

    Not directly comparable

  • AI-Needle

    GPT-5.6 Terra—
    Qwen3.5 397B68.7%
    Source

    Not directly comparable

Knowledge

  • GPQA

    GPT-5.6 Terra92.9%
    Source
    Qwen3.5 397B88.4%
    Source

    GPT-5.6 Terra leads this result

  • GPQA-D

    GPT-5.6 Terra92.9%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • HealthBench Professional

    GPT-5.6 Terra57.7%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • HealthBench Hard

    GPT-5.6 Terra32.7%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • SuperGPQA

    GPT-5.6 Terra—
    Qwen3.5 397B70.4%
    Source

    Not directly comparable

  • MMLU-Pro

    GPT-5.6 Terra—
    Qwen3.5 397B87.8%
    Source

    Not directly comparable

  • MMLU-Redux

    GPT-5.6 Terra—
    Qwen3.5 397B94.9%
    Source

    Not directly comparable

  • C-Eval

    GPT-5.6 Terra—
    Qwen3.5 397B93%
    Source

    Not directly comparable

  • HLE

    GPT-5.6 Terra—
    Qwen3.5 397B28.7%
    Source

    Not directly comparable

Math

  • FrontierMath (legacy)

    GPT-5.6 Terra84.9%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • FrontierMath v2 (Tiers 1-3)

    GPT-5.6 Terra84.900%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • FrontierMath v2 (Tier 4)

    GPT-5.6 Terra68.300%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • AIME26

    GPT-5.6 Terra—
    Qwen3.5 397B93.3%
    Source

    Not directly comparable

  • HMMT Feb 2025

    GPT-5.6 Terra—
    Qwen3.5 397B94.8%
    Source

    Not directly comparable

  • HMMT Nov 2025

    GPT-5.6 Terra—
    Qwen3.5 397B92.7%
    Source

    Not directly comparable

  • HMMT Feb 2026

    GPT-5.6 Terra—
    Qwen3.5 397B87.9%
    Source

    Not directly comparable

  • MMAnswerBench

    GPT-5.6 Terra—
    Qwen3.5 397B80.9%
    Source

    Not directly comparable

Multilingual

  • MMLU-ProX

    GPT-5.6 Terra—
    Qwen3.5 397B84.7%
    Source

    Not directly comparable

  • NOVA-63

    GPT-5.6 Terra—
    Qwen3.5 397B59.1%
    Source

    Not directly comparable

Multimodal

  • MMMU-Pro

    GPT-5.6 Terra80.7%
    Source
    Qwen3.5 397B79%
    Source

    GPT-5.6 Terra leads this result

  • MMMU-Pro w/ Python

    GPT-5.6 Terra82%
    Source
    Qwen3.5 397B—

    Not directly comparable

  • MathVision

    GPT-5.6 Terra—
    Qwen3.5 397B88.6%
    Source

    Not directly comparable

  • CharXiv

    GPT-5.6 Terra—
    Qwen3.5 397B80.8%
    Source

    Not directly comparable

  • VideoMMMU

    GPT-5.6 Terra—
    Qwen3.5 397B84.7%
    Source

    Not directly comparable

  • ScreenSpot Pro

    GPT-5.6 Terra—
    Qwen3.5 397B65.6%
    Source

    Not directly comparable

  • V*

    GPT-5.6 Terra—
    Qwen3.5 397B95.8%
    Source

    Not directly comparable

Instruction following

  • IFEval

    GPT-5.6 Terra—
    Qwen3.5 397B92.6%
    Source

    Not directly comparable

Frequently asked questions

Which is better, GPT-5.6 Terra or Qwen3.5 397B?

GPT-5.6 Terra has the higher public score estimate, 72.29 versus 56.24, 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-5.6 Terra or Qwen3.5 397B?

The current coding 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 is better for agentic tasks, GPT-5.6 Terra or Qwen3.5 397B?

GPT-5.6 Terra leads the like-for-like agentic tasks comparison across 2 shared weighted benchmark rows.

Which costs less, GPT-5.6 Terra or Qwen3.5 397B?

For the stated presets, chat costs $0.008 on GPT-5.6 Terra and $0.0024 on Qwen3.5 397B; repository review costs $0.136 and $0.0408; the cache-heavy agent loop costs $0.2 and $0.168. Qwen3.5 397B does not fit this workload in one request. Qwen3.5 397B has no published cached-input rate, so cached tokens use its listed input rate.

Which has the larger context window, GPT-5.6 Terra or Qwen3.5 397B?

GPT-5.6 Terra has the larger documented context window: 1.05M, compared with 128K.

Related comparisons

Compare API pricing Read the methodology Open the model selector

Last updated August 10, 2026

Watch GPT-5.6 Terra vs Qwen3.5 397B

One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.

Read a sample issue

Join 2,000+ readers.