Updated September 30, 2026. We do not rank this pair: at least one has no public score. 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
At least one model is not scored in the current public ranking lane, so the page does not name an overall quality winner. 2 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
OpenAI
—
Evidence status unavailable
90% interval unavailable
Alibaba Qwen / neutral reproduction
—
Evidence status unavailable
90% interval unavailable
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.
No workload recommendation clears the current evidence threshold.
Use the matched evidence, workload costs, and sourced specifications below instead of treating a point score as a universal answer.
Code generation, repair, and software-engineering tasks
Not enough matched evidence
GPT-5.6 Luna (low reasoning effort) and Raw Qwen3 4B Instruct 2507 direct logits are not ranked on the public lane for coding, so no winner is named for coding.
Tool use, computer use, and multi-step task completion
Not enough matched evidence
GPT-5.6 Luna (low reasoning effort) and Raw Qwen3 4B Instruct 2507 direct logits are not ranked on the public lane for agentic, so no winner is named for agentic.
Prompts that approach the documented context limit
Not enough matched evidence
A complete context comparison is not sourced.
1K fresh input + 500 output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
200K cached + 20K fresh input + 10K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
50K fresh input + 3K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
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.
Not comparable · BenchAlign v5.8
The coding row is not comparable on the public lane: at least one model is not measured or not ranked there.
Coding scores combine specific tasks and setups. Match the editor, harness, and effort to your workflow.
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
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.
Each row shows the public-lane category score for both models: the BenchAlign v5.8 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.
| Category | GPT-5.6 Luna (low reasoning effort) | Raw Qwen3 4B Instruct 2507 direct logits | Basis | Reading |
|---|---|---|---|---|
| Agentic | Not ranked | Not ranked | Not comparableBenchAlign v5.8 lane · 0 vs 0 public rows | Not comparable |
| Coding | Not ranked | Not ranked | Not comparableBenchAlign v5.8 lane · 0 vs 0 public rows | Not comparable |
| Reasoning | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Multimodal | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Knowledge | Not ranked | Not ranked | Not comparableBenchAlign v5.8 lane · 0 vs 0 public rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Instruction following | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Math | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
Ranks count the models scored in each category’s lane, so the agentic, coding, and knowledge denominators (BenchAlign v5.8) 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.
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
GPT-5.6 Luna (low reasoning effort) has no comparable published API token rate. Raw Qwen3 4B Instruct 2507 direct logits has no comparable published API token rate.
50K fresh input + 3K output tokens
GPT-5.6 Luna (low reasoning effort) has no comparable published API token rate. Raw Qwen3 4B Instruct 2507 direct logits has no comparable published API token rate.
200K cached + 20K fresh input + 10K output tokens
GPT-5.6 Luna (low reasoning effort) has no comparable published API token rate. Raw Qwen3 4B Instruct 2507 direct logits has no comparable published API token rate.
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.
GPT-5.6 Luna (low reasoning effort)
Not sourced
Raw Qwen3 4B Instruct 2507 direct logits
Not sourced
GPT-5.6 Luna (low reasoning effort)
Not sourced
Raw Qwen3 4B Instruct 2507 direct logits
Not sourced
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
GPT-5.6 Luna (low reasoning effort)
No comparable hosted API rate
Provider pricingRaw Qwen3 4B Instruct 2507 direct logits
No comparable hosted API rate
Provider pricingGPT-5.6 Luna (low reasoning effort)
Not sourced
Raw Qwen3 4B Instruct 2507 direct logits
Not sourced
GPT-5.6 Luna (low reasoning effort)
Not sourced
Raw Qwen3 4B Instruct 2507 direct logits
Not sourced
GPT-5.6 Luna (low reasoning effort)
Not sourced
Raw Qwen3 4B Instruct 2507 direct logits
Not sourced
GPT-5.6 Luna (low reasoning effort)
Reasoning
Raw Qwen3 4B Instruct 2507 direct logits
Non-Reasoning
GPT-5.6 Luna (low reasoning effort)
Proprietary
Raw Qwen3 4B Instruct 2507 direct logits
Open Weight
GPT-5.6 Luna (low reasoning effort)
Proprietary
Raw Qwen3 4B Instruct 2507 direct logits
Open Weight
GPT-5.6 Luna (low reasoning effort)
Not sourced
Raw Qwen3 4B Instruct 2507 direct logits
Not sourced
At least one model is not scored in the current public ranking lane, so the page does not name an overall quality winner. The page therefore keeps the decision tied to the specific documented workload.
GPT-5.6 Luna (low reasoning effort) and Raw Qwen3 4B Instruct 2507 direct logits are not ranked on the public lane for coding, so no winner is named for coding.
GPT-5.6 Luna (low reasoning effort) and Raw Qwen3 4B Instruct 2507 direct logits are not ranked on the public lane for agentic tasks, so no winner is named for agentic tasks.
Both models do not have comparable published API token rates, so this page does not name a universal price winner.
A complete documented context-window comparison is not available.
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
JevBench 1.4
Shared sourceRaw Qwen3 4B Instruct 2507 direct logits leads this result
JevBench 1.5
Shared sourceRaw Qwen3 4B Instruct 2507 direct logits leads this result
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 30, 2026