Models / GLM-5.3 FlashX Legal
Z.ai
GLM-5.3 FlashX Legal
GLM-5.3 FlashX Legal is strongest at marketing planning; worth a look for simple, high-volume work where cost matters most.
Results 5 Nov 2026 · 88 results · 5 sources
- Price per million tokens
- $0.37 in · $1.25 out
Typical host on OpenRouter, 5 Nov 2026 · compare prices - Speed
- Not measured
- Intelligence Index
- Not measured
- Spring Prompt overall
- 24th of 31 score 34 of 100 · how it works
- Developer
- Z.ai, based in China
Weights not published - Released
- Not recorded
no launch date in our tracker or our sources
1,048,576 tokens of context
Where it stands among the models you could choose
Available models grouped by price, intelligence, speed and developer location. Retired models are excluded.
| Among | Intelligence Index | Spring Prompt overall | Price, blended | Speed |
|---|---|---|---|---|
| Models available today | – | 24th of 31Claude Opus 5.5 | $0.59 blended, 45th percentilecheapest: Mistral Nemo | – |
| Models priced $0.25–1 per million tokens | – | 5th of 6DeepSeek-V4.1-Flash | sets this group | – |
| Models from developers based in China | – | 8th of 10Kimi K3.1 Pro | $0.59 blended, 57th percentilecheapest: Qwen3.5-9B | – |
The name under each rank is the leader of that group; green is the top quarter of the group and red the bottom quarter, by rank. Price is shown as a percentile: the share of the group that costs less, so lower is cheaper. "Sets this group" marks the measure the group is defined by. Spring Prompt overall is our score out of 100 across every source (how it works). Price is the developer's list price (or the typical OpenRouter host where we haven't read one) for three input tokens to one output; intelligence and speed are from Artificial Analysis (data sourced from Artificial Analysis); "similar intelligence" means within 4 points on its Intelligence Index. Developer locations are where each company is based, not where a model is served. Groups under 3 models are not ranked.
Measured cost per task
API cost recorded in our benchmark runs, excluding judging.
| Task | Benchmark | Cost per task |
|---|---|---|
| Product listing | CatalogBench | ≈ $0.005 |
| Deck | DeckBench | ≈ $0.036 |
| Survey plan and analysis | SurveyBench | ≈ $0.069 |
| Game of bullet chess | BulletBench | ≈ $0.006 |
How good is it, and for what?
Each use case ranks the available models its benchmarks measured, then those priced $0.25–1 per million tokens. The bar shows its percentile in that field, best to the right. Open a row for the results behind it.
Product listingsTurning a sparse product feed and photos into listings that can go live
| Benchmark | GLM-5.3 FlashX Legal | Rank among available | Best available |
|---|---|---|---|
| CatalogBenchCatalogBench: reliably publish-ready | 0.0% | 26th of 27 | GPT-6 Astra 75.0% |
| CatalogBenchCatalogBench (sales brief): reliably publish-ready | 0.0% | 22nd of 27 | GPT-6 Astra 69.6% |
Decks from an analysisTurning a finished analysis into a deck you could present as it is
| Benchmark | GLM-5.3 FlashX Legal | Rank among available | Best available |
|---|---|---|---|
| DeckBenchDeckBench: deck rating | 820 | 21st of 28 | GPT-6 Astra 1,606 |
User surveysPlanning a user survey and reading its results without being misled
| Benchmark | GLM-5.3 FlashX Legal | Rank among available | Best available |
|---|---|---|---|
| SurveyBenchSurveyBench: analysis rating | 997 | 8th of 17 | Muse Wonder 1,510 |
Marketing planningPlanning a year of ad spend without overspending
| Benchmark | GLM-5.3 FlashX Legal | Rank among available | Best available |
|---|---|---|---|
| ROASBenchROASBench: overall score | 46.2 | 11th of 28 | GPT-6.1 Terra 58.4 |
Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls
- tau2-bench: has not measured this model.
- Berkeley Function Calling Leaderboard (BFCL) V4: has not measured this model.
- OpenHands Index: has not measured this model.
- Microsoft STATE-Bench: has not measured this model.
- Artificial Analysis: has not measured this model.
- Vending-Bench 2: has not measured this model.
Professional workReal tasks from banking, consulting and law, business documents and freelance projects
- APEX-Agents: has not measured this model.
- GDP.pdf: has not measured this model.
- Remote Labor Index: has not measured this model.
Reasoning and knowledgeHard questions across science, maths and general knowledge
- Artificial Analysis: has not measured this model.
WritingWhat people prefer in blind comparisons, and judged writing quality
- Arena (formerly LMArena): has not measured this model.
- UGI Leaderboard: has not measured this model.
Sticking to the factsSummarising without inventing things, and factual answers
- Vectara Hallucination Leaderboard: has not measured this model.
- Arena (formerly LMArena): has not measured this model.
- SimpleQA Verified (Epoch AI): has not measured this model.
Speed under pressureGood decisions against a real clock (fast chess)
| Benchmark | GLM-5.3 FlashX Legal | Rank among available | Best available |
|---|---|---|---|
| BulletBenchBulletBench Lightning 10+1: ladder Elo | 310 | 12th of 18 | Gemini 3.5 Flash-Lite 892 |
| BulletBenchBulletBench Bullet 60s: ladder Elo | 358 | 16th of 29 | Gemini 3.8 Flash 1,068 |
Nearest alternatives
- The most capable open-weights modelMiMo-V2.6-Pro$0.54 blended · Intelligence Index 46.3
Against GLM-5.3 FlashX Legal at $0.59 blended. "Similar intelligence" means within 4 points or better.
Where to run it
1 host on OpenRouter at the standard service tier, with a median of $0.59 blended per million tokens; quantisation where the host states it: FP8. Cheapest first.
| Host | Quantisation | Input | Output | Blended |
|---|---|---|---|---|
| Z.AI | FP8 | $0.37 | $1.25 | $0.59 |
Prices per million tokens from OpenRouter's public endpoint list, checked 5 Nov 2026.
Against the alternatives
The models you would most likely weigh it against: the leaders of the groups above.
Scroll sideways to see every alternative.
| GLM-5.3 FlashX Legal | DeepSeek-V4.1-Flash highest-scoring overall at $0.25–1 per million tokens | Claude Opus 5.5 highest-scoring overall available today | GLM-5.3 just above overall | Particle-1 just below overall | |
|---|---|---|---|---|---|
| Price per million tokens | $0.59 | $0.52 | $8.00 | $2.15 | $0.16 |
| Tokens a second | – | 369 | 96 | 87 | – |
| Intelligence Index | – | 39.5 | 57.6 | 44.8 | – |
| Spring Prompt overall | 34 | 53 | 84 | 36 | 33 |
| Rank among available models, by use case | |||||
| Product listings | 26th of 27 | 11th | 10th | – | 26th |
| Decks from an analysis | 21st of 28 | – | 4th | – | 15th |
| User surveys | 8th of 17 | – | 3rd | – | 7th |
| Marketing planning | 11th of 28 | – | 7th | 25th | 14th |
| Agents and tool use | – | 28th | 2nd | 13th | – |
| Professional work | – | – | 3rd | 15th | – |
| Reasoning and knowledge | – | 40th | 1st | 29th | – |
| Writing | – | 29th | 5th | 24th | – |
| Sticking to the facts | – | 23rd | 3rd | 52nd | – |
| Speed under pressure | 17th of 31 | – | – | 29th | – |
| Side by side → | Side by side → | Side by side → | |||
Shaded figures are better than GLM-5.3 FlashX Legal's. A dash means no result.
Where it has been measured
Among models available today, in the bottom quarter for product listings.
Not timed: Artificial Analysis, our source for speed, has not tested it
- BulletBench10 results
- CatalogBench30 results
- DeckBench18 results
- ROASBench9 results
- SurveyBench21 results
- APEX-AgentsNot measured
- Arena (formerly LMArena)Not measured
- Artificial AnalysisNot measured
- Berkeley Function Calling Leaderboard (BFCL) V4Not measured
- GDP.pdfNot measured
- Microsoft STATE-BenchNot measured
- OpenHands IndexNot measured
- Remote Labor IndexNot measured
- SimpleQA Verified (Epoch AI)Not measured
- tau2-benchNot measured
- UGI LeaderboardNot measured
- Vectara Hallucination LeaderboardNot measured
- Vending-Bench 2Not measured
Every result
88 published results from 5 sources, each in the source's own units, with the configuration that produced it.
BulletBench · measured by Spring Prompt · 10 results
| Measure | Value | Rank | Configuration | Dated |
|---|---|---|---|---|
| BulletBench Bullet 60s: cost per gameUS dollars, lower is better | $0.0061 | 19 of 44 | GLM-5.3 FlashX Legal (low reasoning) low reasoning |
4 Nov 2026 |
| BulletBench Bullet 60s: games lost on time% of games, lower is better | 37.5% | 19 of 44 | GLM-5.3 FlashX Legal (low reasoning) low reasoning |
4 Nov 2026 |
| BulletBench Bullet 60s: invalid moves% of moves, lower is better | 0.0% | 1 of 44 | GLM-5.3 FlashX Legal (low reasoning) low reasoning |
4 Nov 2026 |
| BulletBench Bullet 60s: ladder Eloladder Elo, higher is better | 358 0–600 |
26 of 44 | GLM-5.3 FlashX Legal (low reasoning) low reasoning |
4 Nov 2026 |
| BulletBench Bullet 60s: median move timemilliseconds, lower is better | 1.4 s | 21 of 44 | GLM-5.3 FlashX Legal (low reasoning) low reasoning |
4 Nov 2026 |
| BulletBench Lightning 10+1: cost per gameUS dollars, lower is better | $0.0035 | 11 of 28 | GLM-5.3 FlashX Legal (low reasoning) low reasoning |
4 Nov 2026 |
| BulletBench Lightning 10+1: games lost on time% of games, lower is better | 62.5% | 20 of 28 | GLM-5.3 FlashX Legal (low reasoning) low reasoning |
4 Nov 2026 |
| BulletBench Lightning 10+1: invalid moves% of moves, lower is better | 0.0% | 1 of 28 | GLM-5.3 FlashX Legal (low reasoning) low reasoning |
4 Nov 2026 |
| BulletBench Lightning 10+1: ladder Eloladder Elo, higher is better | 310 0–641 |
20 of 28 | GLM-5.3 FlashX Legal (low reasoning) low reasoning |
4 Nov 2026 |
| BulletBench Lightning 10+1: median move timemilliseconds, lower is better | 0.9 s | 17 of 28 | GLM-5.3 FlashX Legal (low reasoning) low reasoning |
4 Nov 2026 |
Not shown: Not your cost: prices are those charged on the run date.
CatalogBench · measured by Spring Prompt · 30 results
| Measure | Value | Rank | Configuration | Dated |
|---|---|---|---|---|
| CatalogBench (sales brief): channel rules broken% of products, lower is better | 3.6% | 28 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench (sales brief): content quality% of checks, higher is better | 95.4% | 22 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench (sales brief): failed outputs% of products, lower is better | 1.2% | 24 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench (sales brief): missing UK information% of products, lower is better | 0.0% | 1 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench (sales brief): not findable% of products, lower is better | 0.0% | 1 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench (sales brief): publish-ready listings% of products, higher is better | 0.0% | 27 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench (sales brief): reliably publish-ready% of products, higher is better | 0.0% | 22 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench (sales brief): unsupported claims% of products, lower is better | 79.8% | 22 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench (sales brief): unsupported claimsclaims per product, lower is better | 3.94 | 26 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench (sales brief): wrong attributes% of products, lower is better | 23.2% | 26 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench (sales brief): wrong category or variant% of products, lower is better | 3.0% | 19 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: channel compliance% of products, higher is better | 100.0% | 1 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: channel rules broken% of products, lower is better | 0.0% | 1 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: conflicts caught% of conflicts, higher is better | 100.0% | 1 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: content quality% of checks, higher is better | 96.2% | 20 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: cost per productUS dollars, lower is better | $0.0050 | 10 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: decision accuracy% of decisions, higher is better | 95.1% | 22 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: failed outputs% of products, lower is better | 0.6% | 25 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: field accuracy% of missing fields, higher is better | 94.2% | 11 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: invented values% of filled values, lower is better | 2.2% | 20 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: missing UK information% of products, lower is better | 0.0% | 1 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: not findable% of products, lower is better | 0.0% | 1 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: output tokens per producttokens, lower is better | 3,651 | 6 of 7 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: publish-ready listings% of products, higher is better | 0.0% | 28 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: reliably publish-ready% of products, higher is better | 0.0% | 27 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: time per productseconds, lower is better | 29.4 | 22 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: unsupported claims% of products, lower is better | 27.4% | 19 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: unsupported claimsclaims per product, lower is better | 1.19 | 23 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: wrong attributes% of products, lower is better | 16.7% | 12 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| CatalogBench: wrong category or variant% of products, lower is better | 5.4% | 27 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
Not shown: Not your catalogue: invented products with generated images, so a model's result on your own feed can differ.
DeckBench · measured by Spring Prompt · 18 results
| Measure | Value | Rank | Configuration | Dated |
|---|---|---|---|---|
| DeckBench: accurate decks% of tasks, higher is better | 0.0% | 24 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| DeckBench: caveat dropped% of tasks, lower is better | 50.0% | 29 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| DeckBench: clean layout% of tasks, higher is better | 33.3% | 15 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| DeckBench: cost per deckUS dollars, lower is better | $0.0359 | 10 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| DeckBench: deck ratingrating, higher is better | 820 690–920 |
22 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| DeckBench: design quality% of the maximum, higher is better | 29.3% | 29 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| DeckBench: draft figure quoted% of tasks, lower is better | 83.3% | 29 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| DeckBench: findings missing% of tasks, lower is better | 50.0% | 28 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| DeckBench: head-to-head win rate% of comparisons, higher is better | 47.5% | 19 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| DeckBench: layout defects% of tasks, lower is better | 66.7% | 15 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| DeckBench: misleading metric used% of tasks, lower is better | 50.0% | 29 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| DeckBench: output tokens per decktokens, lower is better | 27,796 | 4 of 8 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| DeckBench: presentable decks% of tasks, higher is better | 0.0% | 13 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| DeckBench: recommendation late or wrong% of tasks, lower is better | 83.3% | 26 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| DeckBench: slides needing work% of tasks, lower is better | 66.7% | 20 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| DeckBench: time per deckseconds, lower is better | 175 | 20 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| DeckBench: unsupported claims% of tasks, lower is better | 66.7% | 28 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| DeckBench: unsupported numbers% of tasks, lower is better | 50.0% | 28 of 29 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
Not shown: Not your analysis or your brand: invented companies, six public tasks, one deck per model per task at the provider's default reasoning setting.
ROASBench · measured by Spring Prompt · 9 results
| Measure | Value | Rank | Configuration | Dated |
|---|---|---|---|---|
| ROASBench: audience scorescore out of 100, higher is better | 63.2 | 14 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| ROASBench: business scorescore out of 100, higher is better | 44.9 | 10 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| ROASBench: consistency scorescore out of 100, higher is better | 29.0 | 12 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| ROASBench: contribution profitsimulated US dollars, higher is better | $695,680 | 7 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| ROASBench: cost of a runUS dollars, lower is better | $0.16 | 14 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| ROASBench: months over budgetmonths of 12, lower is better | 1 | 25 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| ROASBench: overall scorescore out of 100, higher is better | 46.2 | 11 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| ROASBench: planning scorescore out of 100, higher is better | 54.9 | 10 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| ROASBench: return on ad spendprofit per $1 spent, higher is better | $0.55 | 10 of 30 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
Not shown: Not real ad performance: the market is a deterministic simulation of one invented skincare brand.
SurveyBench · measured by Spring Prompt · 21 results
| Measure | Value | Rank | Configuration | Dated |
|---|---|---|---|---|
| SurveyBench: analysis ratingrating, higher is better | 997 875–1,124 |
8 of 18 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: analysis scorepoints of 100, higher is better | 53.4 | 14 of 18 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: cost per taskUS dollars, lower is better | $0.0693 | 9 of 17 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: head-to-head win rate% of comparisons, higher is better | 62.8% | 5 of 18 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: leading or double-barrelled questions% of tasks, lower is better | 44.4% | 9 of 18 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: mistaken next steps% of tasks, lower is better | 44.4% | 11 of 18 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: numbers right% of answers, higher is better | 94.9% | 13 of 18 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: objectives unanswerable% of tasks, lower is better | 22.2% | 14 of 18 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: output tokens per tasktokens, lower is better | 51,553 | 4 of 7 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: precision ignored% of tasks, lower is better | 22.2% | 10 of 18 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: research scorepoints of 100, higher is better | 59.2 | 13 of 18 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: sound research% of tasks, higher is better | 11.1% | 7 of 18 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: stakeholder's view presumed% of tasks, lower is better | 22.2% | 10 of 18 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: survey plan scorepoints of 100, higher is better | 72.7 | 15 of 18 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: survey structure% of tasks, lower is better | 44.4% | 15 of 18 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: time per taskseconds, lower is better | 435 | 15 of 17 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: traps handled% of traps, higher is better | 83.0% | 10 of 18 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: traps missed% of tasks, lower is better | 66.7% | 8 of 18 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: unsupported findings% of tasks, lower is better | 77.8% | 9 of 18 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: wrong numbers% of tasks, lower is better | 33.3% | 10 of 18 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
| SurveyBench: wrong recommendation% of tasks, lower is better | 55.6% | 14 of 18 | GLM-5.3 FlashX Legal provider default reasoning |
4 Nov 2026 |
Not shown: Not your customers or your data: invented organisations with simulated respondents, nine public tasks, one run per model per task at the provider's default reasoning setting.