Agentic work
Tool use, computer use, and multi-step task completion
GPT-5.6 Terra
GPT-5.6 Terra leads on the public agentic lane, 59.5 to 20.1, with Supported evidence for both models and non-overlapping 90% intervals.
Confidence: stronger
GPT-5.6 Terra vs Mistral Medium 3.5 128B
Updated September 24, 2026. Rank says GPT-5.6 Terra 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-5.6 Terra has the higher public score, 72.58 versus 36.05, and the 90% score intervals do not overlap. 4 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
- Shared results
- 4
- GPT-5.6 Terra only
- 28
- Mistral Medium 3.5 128B only
- 3
- Like-for-like categories
- 2 / 8
Supported: GPT-5.6 Terra · Estimated: Mistral Medium 3.5 128BHow 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-5.6 Terra
GPT-5.6 Terra has the larger documented context window.
Confidence: documentedChat turn cost
1K fresh input + 500 output tokens
Mistral Medium 3.5 128B
Mistral Medium 3.5 128B 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
GPT-5.6 Terra
GPT-5.6 Terra has the lower estimated token cost for this stated workload. Mistral Medium 3.5 128B 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
Mistral Medium 3.5 128B
Mistral Medium 3.5 128B has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-ratesCoding work
Code generation, repair, and software-engineering tasks
Not enough matched evidence
Mistral Medium 3.5 128B is scored on Estimated evidence for coding, 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.
64.7GPT-5.6 Terra26.9Mistral Medium 3.5 128B
Directional only · BenchAlign v5.7
GPT-5.6 Terra scores higher, but at least one score rests on Estimated evidence or a different benchmark set. Directional only, no winner.
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.
28 GPT-5.6 Terra only
4 Shared
3 Mistral Medium 3.5 128B only
2 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.
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.
MMLU-Pro (Vals)Knowledge
Normalized gap 11.4GPT-5.6 Terra:86.7%Mistral Medium 3.5 128B:75.3%
Bars run 0–100 on each benchmark’s normalized display scale
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.
Agentic
Like-for-like- GPT-5.6 Terra
- 59.5
- Supported · #18/105
- Mistral Medium 3.5 128B
- 20.1
- Supported · #86/105
- Basis
- BenchAlign v5.7 lane · 9 vs 3 public rows
- Reading
- GPT-5.6 Terra leads
Knowledge
Like-for-like- GPT-5.6 Terra
- 71.1
- Supported · #9/158
- Mistral Medium 3.5 128B
- 32.1
- Supported · #120/158
- Basis
- BenchAlign v5.7 lane · 8 vs 2 public rows
- Reading
- GPT-5.6 Terra leads
Coding
Directional only- GPT-5.6 Terra
- 64.7
- Supported · #8/135
- Mistral Medium 3.5 128B
- 26.9
- Estimated · #94/135
- Basis
- BenchAlign v5.7 lane · 8 vs 2 public rows
- Reading
- Directional only
Instruction following
Directional only- GPT-5.6 Terra
- 85.7
- #35/124
- Mistral Medium 3.5 128B
- 82.6
- #49/124
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Directional only
Reasoning
Not comparable- GPT-5.6 Terra
- 65.4
- #12/19
- Mistral Medium 3.5 128B
- 69.7
- Unranked · 2 rankable rows
- Basis
- Provisional lane · 2 vs 0 weighted rows
- Reading
- Not comparable
Multimodal
Not comparable- GPT-5.6 Terra
- 77.2
- #16/50
- Mistral Medium 3.5 128B
- 55.7
- Unranked · 1 rankable row
- Basis
- Provisional lane · 1 vs 0 weighted rows
- Reading
- Not comparable
Multilingual
Not comparable- GPT-5.6 Terra
- Not ranked
- Mistral Medium 3.5 128B
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Math
Not comparable- GPT-5.6 Terra
- 96.8
- Unranked · 3 rankable rows
- Mistral Medium 3.5 128B
- Not ranked
- Basis
- Provisional lane · 2 vs 0 weighted rows
- Reading
- Not comparable
| Category | GPT-5.6 Terra | Mistral Medium 3.5 128B | Basis | Reading |
|---|---|---|---|---|
| Agentic | 59.5Supported · #18/105 | 20.1Supported · #86/105 | Like-for-likeBenchAlign v5.7 lane · 9 vs 3 public rows | GPT-5.6 Terra leads |
| Knowledge | 71.1Supported · #9/158 | 32.1Supported · #120/158 | Like-for-likeBenchAlign v5.7 lane · 8 vs 2 public rows | GPT-5.6 Terra leads |
| Coding | 64.7Supported · #8/135 | 26.9Estimated · #94/135 | Directional onlyBenchAlign v5.7 lane · 8 vs 2 public rows | Directional only |
| Instruction following | 85.7#35/124 | 82.6#49/124 | Directional onlyProvisional lane · 0 vs 0 weighted rows | Directional only |
| Reasoning | 65.4#12/19 | 69.7Unranked · 2 rankable rows | Not comparableProvisional lane · 2 vs 0 weighted rows | Not comparable |
| Multimodal | 77.2#16/50 | 55.7Unranked · 1 rankable row | Not comparableProvisional lane · 1 vs 0 weighted rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Math | 96.8Unranked · 3 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-5.6 Terra
- $0.008
- Fits in one request
- Mistral Medium 3.5 128B
- $0.00525
- Fits in one request
Mistral Medium 3.5 128B 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
- Mistral Medium 3.5 128B
- $0.0975
- Fits in one request
Mistral Medium 3.5 128B 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
- Mistral Medium 3.5 128B
- $0.405
- Fits in one request
- Cached input priced at the published list-input rate
GPT-5.6 Terra has the lower modeled cost
Mistral Medium 3.5 128B 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-5.6 TerraMistral Medium 3.5 128B
Context window
Maximum documented context; output-token limits may be lower.
GPT-5.6 Terra
1.05M
OpenAI model catalogMistral Medium 3.5 128B
256K
API model ID
GPT-5.6 Terra
gpt-5.6-terra
OpenAI model catalogMistral Medium 3.5 128B
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
GPT-5.6 Terra
$0.2 per 1M cached input tokens
OpenAI pricingMistral Medium 3.5 128B
Not published
Documented inputs
GPT-5.6 Terra
text, image
OpenAI model catalogMistral Medium 3.5 128B
Not sourced
Documented outputs
GPT-5.6 Terra
Mistral Medium 3.5 128B
Not sourced
Provider availability
GPT-5.6 Terra
Generally Available · OpenAI Responses API
OpenAI model catalogMistral Medium 3.5 128B
Not sourced
Reasoning profile
GPT-5.6 Terra
Reasoning
Mistral Medium 3.5 128B
Reasoning
Weight access
GPT-5.6 Terra
Proprietary
Mistral Medium 3.5 128B
Open Weight
License
GPT-5.6 Terra
Proprietary
Mistral Medium 3.5 128B
Open Weight
Release date
GPT-5.6 Terra
2026-07-09
Mistral Medium 3.5 128B
2026-04-29
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, 72.58 versus 36.05, and the 90% score intervals do not overlap.
- Workload cost
- Repository review: $0.136 vs $0.0975. Cache-heavy agent loop: $0.2 vs $0.405.
- 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.
Questions
Which is better, GPT-5.6 Terra or Mistral Medium 3.5 128B?
GPT-5.6 Terra has the higher public score, 72.58 versus 36.05, 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-5.6 Terra or Mistral Medium 3.5 128B?
GPT-5.6 Terra scores higher for coding on the public lane, 64.7 to 26.9. Mistral Medium 3.5 128B is scored on Estimated evidence for coding, so the reading is directional rather than like-for-like. BenchLM does not name a winner for coding from a directional reading; 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 Mistral Medium 3.5 128B?
GPT-5.6 Terra leads the public agentic tasks lane, 59.5 to 20.1, with Supported evidence for both models and non-overlapping 90% intervals.
Which costs less, GPT-5.6 Terra or Mistral Medium 3.5 128B?
For the stated presets, chat costs $0.008 on GPT-5.6 Terra and $0.00525 on Mistral Medium 3.5 128B; repository review costs $0.136 and $0.0975; the cache-heavy agent loop costs $0.2 and $0.405. Mistral Medium 3.5 128B 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 Mistral Medium 3.5 128B?
GPT-5.6 Terra has the larger documented context window: 1.05M, compared with 256K.
Benchmark evidence
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Browse raw public benchmark evidence35 rows
Agentic
Terminal-Bench 3.0
GPT-5.6 Terra20.8%SourceMistral Medium 3.5 128B—Not directly comparable
Terminal-Bench 2.1
GPT-5.6 Terra87.4%SourceMistral Medium 3.5 128B—Not directly comparable
BrowseComp
GPT-5.6 Terra87.5%SourceMistral Medium 3.5 128B—Not directly comparable
OSWorld 2.0
GPT-5.6 Terra50.2%SourceMistral Medium 3.5 128B—Not directly comparable
CyberGym
GPT-5.6 Terra81.8%SourceMistral Medium 3.5 128B—Not directly comparable
ExploitGym
GPT-5.6 Terra23.2%SourceMistral Medium 3.5 128B—Not directly comparable
Toolathlon
GPT-5.6 Terra53.1%SourceMistral Medium 3.5 128B—Not directly comparable
Terminal-Bench 2.1 (Vals)
GPT-5.6 Terra77.5%SourceMistral Medium 3.5 128B39.0%SourceGPT-5.6 Terra leads this result
ApprenticeBench
GPT-5.6 Terra16%SourceMistral Medium 3.5 128B—Not directly comparable
τ³-bench results
GPT-5.6 Terra—Mistral Medium 3.5 128B91.4%SourceNot directly comparable
Gert Labs
GPT-5.6 Terra—Mistral Medium 3.5 128B39.10%SourceNot directly comparable
Coding
SWE-bench Pro
GPT-5.6 Terra63.4%SourceMistral Medium 3.5 128B—Not directly comparable
Terminal-Bench 2.1
GPT-5.6 Terra87.4%SourceMistral Medium 3.5 128B—Not directly comparable
DeepSWE
GPT-5.6 Terra69.6%SourceMistral Medium 3.5 128B—Not directly comparable
FrontierCode 1.1 Extended
GPT-5.6 Terra55.8%SourceMistral Medium 3.5 128B—Not directly comparable
cursorBench32
GPT-5.6 Terra64.9%SourceMistral Medium 3.5 128B—Not directly comparable
VulcanBench v3
GPT-5.6 Terra87.0%SourceMistral Medium 3.5 128B—Not directly comparable
LiveCodeBench (Vals)
GPT-5.6 Terra85.9%SourceMistral Medium 3.5 128B—Not directly comparable
SWE-bench (Vals)
GPT-5.6 Terra95.4%SourceMistral Medium 3.5 128B66.4%SourceGPT-5.6 Terra leads this result
SWE-bench Verified
GPT-5.6 Terra—Mistral Medium 3.5 128B77.6%SourceNot directly comparable
Reasoning
Multimodal
Knowledge
GPQA
GPT-5.6 Terra92.9%SourceMistral Medium 3.5 128B—Not directly comparable
GPQA-D
GPT-5.6 Terra92.9%SourceMistral Medium 3.5 128B—Not directly comparable
HLE-Verified
GPT-5.6 Terra51.1%SourceMistral Medium 3.5 128B—Not directly comparable
LABBench2
GPT-5.6 Terra81.2%SourceMistral Medium 3.5 128B—Not directly comparable
HealthBench Professional
GPT-5.6 Terra57.7%SourceMistral Medium 3.5 128B—Not directly comparable
HealthBench Hard
GPT-5.6 Terra32.7%SourceMistral Medium 3.5 128B—Not directly comparable
GPQA Diamond (Vals)
GPT-5.6 Terra90.9%SourceMistral Medium 3.5 128B34.8%SourceGPT-5.6 Terra leads this result
MMLU-Pro (Vals)
GPT-5.6 Terra86.7%SourceMistral Medium 3.5 128B75.3%SourceGPT-5.6 Terra leads this result
Math
35 public results · 4 shared
Watch GPT-5.6 Terra vs Mistral Medium 3.5 128B
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 24, 2026